Semantics and Mental Representation in Aristotle s Peri Hermeneias. Rodmon Cedric King. Submitted in Partial Fulfillment. of the

Transcription

1 Semantics and Mental Representation in Aristotle s Peri Hermeneias by Rodmon Cedric King Submitted in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy Supervised by Professor Deborah Modrak Department of Philosophy Arts, Sciences and Engineering School of Arts and Sciences University of Rochester Rochester, New York 2012

2 ii For Edith Virginia King We must strive to make each generation better than the last Joshua, Josiah, Jeremiah, and Jared King May you live lives worthy of the excellence within you

3 iii Curriculum Vitae Rodmon Cedric King was born in Rochester, New York on October 24, He attended Roberts Wesleyan College and graduated with a Bachelor of Arts degree in Religion and Philosophy in He came to the University of Rochester in the fall of 2000 and began graduate studies in philosophy. He received a University Fellowship in 2000, a teaching assistantship in 2001, and a Writing Fellowship in He received a Consortium for Faculty Diversity Fellowship in He was nominated for the Edward Peck Curtis Award for Teaching Excellence in He received the Edward Peck Curtis Award for Teaching Excellence in He pursued his research under the direction of Professor Deborah Modrak and received the Master of Arts degree from the University of Rochester in He is currently an Instructor of Philosophy at Hobart & William Smith Colleges. He is the father of four amazing sons.

4 iv Acknowledgements My deepest and most sincere thanks to my committee: Deborah Modrak, Randall Curren, and Edward Wierenga. They have provided me crucial support and assistance in my research and writing. My advisor Deborah Modrak has been particularly supportive of my efforts with countless hours of vital feedback, and encouragement. This thesis would not have moved from potentiality to actuality without her sound guidance. I could not have wished for a more dedicated mentor and advisor. I thank the current and former members of the University of Rochester Philosophy Department for the hours they spent mentoring and inspiring me. In particular, I thank Amy Bray for everything she does (often without recognition) so that work such as this sees the light of day. In addition, I thank Gabriele Uzquiano, David Braun, Richard Feldman, John Bennett, Ralf Merebote, and Earl Conee who have each played a vital role in my intellectual development. I also thank my former student colleagues Jacqueline Augustine, Stefan Forrester, Pat Kenny, Gregory Jansen, Todd Long, and Eva Cadavid. Their support and friendship helped to sustain me through this process. In particular, I am grateful to Eva Cadavid for simply being the wonderful friend and person she is. I also thank the staff and librarians at Rush-Rhees library. I am especially grateful to Eileen Daly for her vital research assistance. I extend thanks to my colleagues and peers at Hobart & William Smith Colleges: Scott Brophy, Steven Lee, Karen Frost-Arnold, Eric Barnes, Greg Frost-

5 v Arnold, Carol Oberbrunner, and Eugen Baer. In particular, I thank Scott Brophy for endless support and advocacy. I am also grateful to President Mark Gearan, Interim Provost Pat McGuire, and former Provost Teresa Amott. In addition, I thank Pat Cool, Kelly Switzer, Tina (Phillip) Smaldone, Judy Mahoney-Benzer, and J.J. Smaldone for keeping me grounded. I express my appreciation and gratitude to the Consortium for Faculty Diversity. My thanks go to Paul Livermore and David Basinger at Roberts Wesleyan College for restoring my love of learning and providing me the skills necessary for graduate studies. I am grateful to my family for their prayers, support, and encouragement. I especially want to thank my parents Roy and Lucille King for instilling in me the strength to pursue my dreams. Also, I am grateful to my sons Joshua, Josiah, Jeremiah, and Jared for their patience throughout this process. In addition, I thank my siblings Rashid Muhammad, Lisa Harris, and Philip King for their love and support. I also want to thank my extended family for their prayers and encouragement throughout this process. Lastly and most importantly, I want to thank Sarah Berry without whose love and support I would never have been able to complete this thesis. I cannot imagine a more caring and supportive life partner. Each and everyday I am thankful that we found each other.

6 vi Abstract This dissertation is a careful examination of the first six chapters of Peri Hermeneias. This text contains what many consider to be Aristotle s most explicit claims concerning language and meaning. In the first half of the last century, Aristotle s claims in Peri Hermeneias came under withering criticism. Recently, Aristotelian scholarship has entered an exciting period as it has been argued that his linguistic claims in Peri Hermeneias are deep, sophisticated, and defensible. This recent work has successfully debunked various misinterpretations of Aristotle s claims in Peri Hermeneias. With significant amounts of the turbid air surrounding Peri Hermeneias cleared, the path is open for renewed investigations into and interpretations of Aristotle s claims. My intervention on one hand is to critically evaluate the interpretations of Aristotle s claims in Peri Hermeneias 1-6. Simultaneously, on the other hand, my intervention is to present a new interpretation of Aristotle s concept of signification and explore its implications. The first chapter of this thesis is a focused discussion of Plato s analysis of conventionalism and naturalism in Cratylus and the way in which this analysis sets the stage for Aristotle s discussion of language in Peri Hermeneias. The second chapter critically examines the issues with which the early Greek commentators were concerned. Specifically, I examine their interpretations of affections of the soul (παθήµατα), sign, symbol, likeness, and the connection between Aristotle s claims in Peri Hermeneias and his other works. In the third chapter, I argue for

7 vii specific interpretations of the sign, symbol, and likeness relations based upon a balanced view of the philological and philosophical evidence. Also, I argue for φαντάσµατα as the affections of the soul (παθήµατα) referenced by Aristotle in Peri Hermeneias 1, and I connect this interpretation to Aristotle s work in the Categories and his account of perception. In the fourth chapter, I argue for an interpretation of Aristotle s concept of signification that centers on the representational power of φαντάσµατα and I apply this interpretation to the linguistic phenomena Aristotle discusses in Peri Hermeneias 1-6. Lastly, in the fifth chapter, I summarize my interventions and suggest possible applications of my interpretive work.

8 viii Table of Contents Chapter 1 Platonic Background Outline and Summary of the Cratylus The Correctness of Names and the Truth-Value of Statements The Correctness of Names and Conventionalism Language and Ontology The Tool Analogy The Memetic/Onomatopoetic Theory Summary and Findings 40 Chapter 2 The Early Greek Commentary Tradition and Peri Hermeneias Peri Hermeneias The Early Greek Commentary Tradition The Schema of Chapter One of Peri Hermeneias Φωναί (Spoken Sounds): Names (ὄνοµατα) and Verbs (ῤῆµατα) Sentences and Affirmations Findings and conclusions 73 Chapter 3 Modern Commentators and Peri Hermeneias Sign and Symbol Likeness (ὀµοίωµατα) and Affections in the Soul (παθήµατα) 85

9 ix 3.3 Real Things 113 Chapter 4 Signification and Semantic Import The relations in Peri Hermeneias Likeness and Signification Likeness Signification Signification and Peri Hermeneias Ὄνοµατα (Names) Ρῆµατα (Verbs) Indefinite Names and Verbs Λόγοs (Sentence) 150 Chapter 5 Conclusions and Implications 154 Bibliography 163

10 x List of Tables Table Title Table 1 Tool Analogy 27 Table 2 Schema for Aristotle s Categories 117 Table 3 Relations in Peri Hermeneias 16a

11 xi List of Figures Figure Title Figure 1 Schema for Vocal Sounds 58 Figure 2 Statement-Making Sentences 72 Figure 3 Aristotelian Taxonomy of Sound 88 Figure 4 Object of Perception and Phantasia 100 Figure 5 The process of perception of color 110 Figure 6 The process for thoughts of universals 111 Figure 7 Schema for the Category of Quality 118 Figure 8 Schema for the Category of Substance 119 Figure 9 Statement-Making Sentences 151

12 1 CHAPTER ONE - Platonic Background Introduction I begin my investigation of Aristotle s thought in Peri Hermeneias by exploring Plato s Cratylus because it presents key aspects of Plato s thought on language and meaning. 1 Plato's discussion, not surprisingly, sets the stage for Aristotle's discussion of language in Peri Hermeneias. 2 This chapter is dedicated to the discussion of three important claims found tacitly and/or explicitly in the Cratylus that connect to Aristotle's work in Peri Hermeneias: (Α) The correct account of names may explain the truth-value assigned to names and statements In one of the most controversial passages in the dialogue, Socrates is able to convince Hermogenes that there are true and false statements and that all the parts 1 Specifically, this dialogue explores the relationship between linguistic symbols and the things they represent. 2 I am merely pointing out that Aristotle s investigation can be viewed as being indebted to Plato s work on language. This is not to make the tempting, and controversial claims that the Cratylus was unequivocally Aristotle's intended target when he wrote Peri Hermeneias. What is clear is that the central problem of the Cratylus (i.e. the problem of giving an adequate account of meaning) sets the stage for Aristotle's exploration of this and other issues.

13 2 of a true statement are true (385b-d). 3 Hermogenes then concedes that it is possible to name things truly or falsely. I believe that there are two important things to take from Plato's discussion of the truth-value of names and statements in the Cratylus. First, it seems that Plato is endorsing the not very surprising conclusion that the correct account of names will allow for the differentiation between proper and improper applications of a name. The truth or falsity of a name lies in its ability to refer to the nominatum and its ability to accurately describe the nominatum. Second, Plato appears to invoke a view where the names are true and false in the same way that statements are true and false. This set of claims as well as Plato s other claims in the Cratylus regarding truth-values are analyzed in section 1.2 of this chapter. In chapter four, I show that Aristotle also discusses the implications of his account of language for the truth-value of statements. (Β) The correct account of names could contain a (possibly necessary) conventional element. Hermogenes claims, No name belongs to a particular thing by nature, but only because of the rules and usage of those who establish the usage and call it by that name (384b4-6). During his conversation with Cratylus, Socrates points out 3 This passage has been the subject of much discussion. Commentators have disagreed about the structure of the argument, its purpose, and its place in the overall makeup of the Cratylus. See footnote 7 for further detail on these moves within the literature on the Cratylus.

14 3 that it appears that the correctness of names is not only a matter of names being like the things they name, but usage/convention also has a role (435b1-c2). Socrates claims that it is usage/convention that allows both like and unlike letters to express things (435b2). In section 1.3 of this chapter I examine the conventionalist accounts in the Cratylus. In Chapter Four, I show that there is a conventional element in Aristotle s thought on language. At Peri Hermeneias 16a20 Aristotle says, "A name is spoken sound significant by convention." Later he also says, "I say 'by convention' because no name is a name naturally but only when it has become a symbol" (16a27). I interrogate the meaning of these claims and what commitments they entail for Aristotle s account of language. (Γ) The correct account of names will explicate the connection between ontology and language. A major part of the Cratylus is devoted to investigating the natural correctness of names (i.e., how names embody the things that are). 4 Among the claims examined by Socrates and his interlocutors are the statements that names are images (or likenesses) of things and that names are tools for dividing being. One of the challenges that Aristotle inherits from Plato is the task of supplying an adequate account of the relation between words and language independent reality. These claims are examined in section 1.4 of this chapter. 4 While this topic is explored in the conversation with Hermogenes and the conversation with Cratylus, it is the central topic in the conversation with Cratylus.

15 4 1.1 Outline and Summary of the Cratylus Introduction When read for the first time, the discussion in the Cratylus appears to be solely an etymological examination of the correctness of names. 5 However, as I demonstrate there is much more going on in the dialogue in the sections that follow. Socrates and his interlocutors attempt to discover a criterion for correct application of names (e.g., societal convention or an ontological connection/relation between names and the things named). In the dialogue, two distinct accounts of the correctness of names are explored. One account proposed by Hermogenes is that the correctness of names is established purely by convention. The first half of the dialogue is dedicated 5 This phrase sounds very odd to modern ears. Palmer (1989) provides an excellent explanation of this phrase. Palmer suggests that there are two ways in which this phrase is used in the Cratylus: "(1) the way in which it is decided that a particular phonetic unit or collocation of marks constitutes a name or (2) the basis for deciding what names there ought to be." (27) Both the conventionalist and naturalist views include both of these conceptions of the correctness of names. In terms of (1) a conventionalist view holds that "humans decide what sounds are to function as names." (27) A naturalist view of this type "holds that sounds are names because they are phonetic imitations of the things they designate." In terms of (2) a conventionalist view holds that "the distinctions we make among things exist only by convention, and that names are merely phonetic units chosen to call attention to these distinctions." (28) A naturalist view of this type holds that "the structure of reality itself is the ultimate criterion both for determining correct reference of particular names and for determining what names there ought to be" (28).

16 5 to the exploration and evaluation of Hermogenes view. The other account, advanced by Cratylus, is that the correct name belongs to each thing according to the nature of the thing named. The second half of the dialogue is concerned with investigating Cratylus account and its implications. By the end of the Cratylus, Plato rejects the two main candidate accounts for naming: conventionalism and naturalism. In Socrates conversation with Hermogenes, conventionalism is rejected because words are tools for dividing reality and the way in which words perform this division is not merely a matter of personal or societal convention. In Socrates conversation with Cratylus, naturalism is rejected on the grounds that names cannot provide us with a non-arbitrary way to decide between the competing ontological accounts for a basis of naming and consequently an adequate account cannot be given of the connection between a name and an object it represents.

17 6 1.2 The Correctness of Names and the Truth-Value of Statements Now that I have reviewed the Cratylus in outline I can focus on how some of the major claims in this dialogue set the stage for Aristotle's investigations in Peri Hermeneias. First let us examine the first claim: (Α) The correct account of names will explain the truth-value assigned to names and statements. This claim, while not explicitly stated either in Socrates conversation with Hermogenes or his conversation with Cratylus, is central to Plato's arguments in the dialogue. The key issue in this dialogue is the correctness of names. As I reveal, correct names are names that are true in that they refer to the nominatum and make evident the essence of the thing named. Also, I show that Plato s discussion of true and false names plays a central role in the rejection of conventionalism and naturalism as the sole criterion (respectively) for correctness of names. The concepts of truth and falsity first appear after Socrates has finished an analogical argument for the conclusion that naming must be performed in a way

18 7 similar to other actions. 6 Socrates then presents the following argument to Hermogenes (385c5ff): 7 (1) There is such a thing as speaking the truth and speaking falsehood (385b2-5) (2) There is such a thing as true statements and false statements (385b-5-6) (3) Statements which say things as they are are true and statements which says things as they are not are false (385b7-9) (4) Therefore, it is possible to say, in a statement, that which is and that which is not (385b10-12) (5) If a true statement is true as a whole, then its parts are also true (385b13- c5) (6) If a false statement is false as a whole, then at least one part is false (7) The smallest part of a statement is a name (385c7-8) (8) Therefore it is possible to say true and false names (385c10-17) 6 See page 27 of this dissertation. 7 I follow Schofield (1972) and transfer 385b2-d1 to follow 387c5 in order to preserve the logical progression of Socrates' argument. For an excellent discussion of defenses and objections to this move see Baxter (1992), 32-34; Sedley (2003), 10-13; Fine (1977). This change is also adopted by Cooper (1997) and Reeve (1998).

19 8 This argument has been the subject of much debate. 8 T.M Baxter (1992) states that this argument is "quite possibly the most discussed passage in the whole dialogue." 9 Many commentators have objected to this argument. Most prominently, R. Robinson (1969) claims that there are two major errors in this argument. First the argument commits the fallacy of division in premises (5) and (6). 10 Secondly, the argument wrongly ascribes descriptive content to names and in so doing conflates naming and stating. 11 Robinson himself does not spend much time developing the first criticism. However, a quick reading of the argument shows that Plato seems to be improperly arguing from a whole to parts at premises (4) or (5) to allow that whatever truth-value a statement has, each of its parts has also. But this kind of reasoning is clearly fallacious. The fact that something can be truly said of a whole (in this case a statement) does not make the case that the same thing can be said of any part of that whole (in this case a name). 8 Although I will examine various positions regarding this argument in the literature, I do not want to lose sight of the fact that my interest in this debate in the literature is solely to show how the account of truth and falsity in the Cratylus sets the stage for further discussion of the topic by Aristotle in Peri Hermeneias. I do not wish to prove that this argument is valid. I only wish to make the case that this argument is part of an intricate account of truth and falsity with respect to names and statements and elements of this account can be seen in Aristotle's work in Peri Hermeneias. 9 Baxter (1992), Robinson (1969), Robinson (1969), 130.

20 9 Robinson (1969) devotes more space to the second objection. He repeatedly states that the only function a name has is to refer to a thing or a set of things. 12 This is contrasted with the function of a statement, which is to provide a description of a thing or state of affairs. 13 Robinson (1969) goes on to claim "names are neither true nor false because they do not assert or describe, but name or refer." 14 According to Robinson, statements are true or false because they do exactly what names cannot (i.e. describe and assert). Hence, the argument is wrong to claim that names can be true or false, because doing so conflates the separate functions of names and statements. 15 The argument in effect makes names into little sentences or descriptions. In support of his reading of this argument, Robinson cites the account of truth and falsity in the Sophist. There the distinction between stating and naming is clearly made. Stating goes beyond naming in that it reveals (δῆλοι) something by weaving together names (ὄνοµατα) and verbs (ῥῆµατα) to form sentences (λόγοι). 16 A distinction here is made between names, whose function is to refer, Op. Cit., 130, 132, and Op. Cit., Op. Cit., Ibid. 16 Sophist 262d. 17 At 262a4 a name is described as a "kind of spoken sign that's applied to things that perform actions."

21 10 and statements, which can be assigned truth-values. It would appear that Plato, in the Sophist, rejects the earlier argument put forth in the Cratylus. Fine (1977) argues that Robinson's analysis of this argument can be resisted. Fine suggests that Robinson has misunderstood the claims made in the argument. Fine (1977) goes on to claim that Plato is not committed to names being "little sentences, or true and false just as sentences are." 18 When understood properly, Plato's use of 'true' and 'false' does not make this argument susceptible to Robinson's second criticism. This, however, does not mean that this argument is valid. 19 To address Robinson's claim that names only refer, Fine claims that Plato attaches descriptive content to names and therefore is not committed here to the view that names only refer. 20 To support her position, Fine (1977) cites evidence from the etymological sections of the dialogue. 'Astyanax' is the proper name for Hector's son (392e) not only because it refers to him, but also because it is descriptively accurate. 21 In response to Robinson's other claim that names are neither true nor false, Fine cites the discussion of false names in the conversation with Cratylus (429b- 432e). Cratylus denies that speaking falsely is possible (429c1ff). At 429d4, Cratylus asks "doesn't speaking falsely consist in not saying things that are?" This clear statement of the third premise of the argument goes unnoticed by both Fine 18 Fine (1977), Op Cit., Robinson (1969), 293 and Op. Cit., 297.

22 11 and Robinson, but it does seem to show a connection between the argument at 385c5ff 22 and the discussion of falsity in the conversation with Cratylus. Cratylus' position seems to be motivated by the argument at 385c5ff. As Baxter (1992) points out, the argument at 385dff demands that πρῶτον ὄνοµατα (primary names) have truth value and all the πρῶτον ὄνοµατα of any ὕστερον ὄνοµα (complex name) must be true if that primary name is to be true. 23 The upshot of this is that "either a complex name is 'completely' true down to each πρϖτον ὄνοµα, or it does not refer to (describe) that object at all through the lack or superfluity of even one πρϖτον ὄνοµα." 24 Hence, at 429b-c, Cratylus' statement that names are given correctly or not at all. Fine does connect Plato's response to Cratylus at 430aff with the argument at 385c5ff. 25 Plato uses an analogy of pictures to convince Cratylus that just as there can be misassignment of pictures there can be misassignment of names. 26 Such misassignment constitutes speaking falsely. 27 If Plato means that names can be true or false in the terms of misassignment then, according to Fine, the argument at 385c5ff does not make names true or false in the same way that 22 See section of this thesis. 23 Baxter (1992), 36. This follows from applying the account of primary names from 422b2ff to the argument at 385c5ff. 24 Ibid. 25 Fine (1977), On page 298, Fine states that The tall blonde man with one black shoe is a description but not a sentence a-d discussed on page 13 of this dissertation b1ff discussed on page 13 of this dissertation.

23 12 sentences are. On this view, names "are true or false of things and can rightly or wrongly be applied to them." 28 One might still be concerned that the descriptive content associated with a name does make names into little sentences as Robinson claims. Fine correctly responds to this by pointing out that not all descriptions are sentences. 29 Therefore, names could have descriptive content without being little sentences. This analogy also makes a connection to the account of truth and falsity in the Sophist that Robinson cited. Socrates, at 431b7ff, states that the same account holds for verbs (i.e. misassignment of verbs in false statements) and it also holds for statements because statements are combination of names and verbs. 30 In the Sophist at 263b5-7, it is stated that false sentences speak of things different than those that are. Specifically, the sentence Theatetus flies says that things are real of Theatetus that in fact are not (263a7-b7). In other words the verb flies is misassigned to Theatetus. These passages are consistent in that they both connect the falsity of a statement with misassignment of one of the constituents of the statement. 28 Fine (1977), Op. Cit., 298. Fine points out that some descriptions (e.g., the tall blond man with one black shoe) are not sentences. Fine states that Robinson has not supplied sufficient information to establish that Plato assimilates words to sentences. 30 This would seem to contradict Fine claim that Plato does not hold that names and statements are true or false in the same way. It is not necessary to defend Fine's on this point. The overall purpose here is to show that there is a consistent intricate account of the truth and falsity of names and statements.

24 13 The upshot of all this is that the account of true and false names and statements in the Cratylus is more developed than Robinson acknowledges. As Baxter (1992) points out, the argument at 385c5ff can be seen as "a bare sketch of a prescriptive theory" for language and 429b-432e can be seen as developing this sketch. 31 Verbs are distinguished from names, and there is a shift away from a singular focus on names. The account gives a ground up explanation (from primary elements to complexes) as to why words and statements are true. In the next section of this chapter I indicate that this account leads to the reintroduction of convention as an element in the correct account of names Op. Cit., In later chapters we will see that Aristotle, in Peri Hermeneias, spends a good amount of time developing an account of names and affirmations with a specific focus on the sub-class of affirmations that have truth values.

25 The Correctness of Names and Convention I can now move on to examine (Β): The correct account of names could contain a (possibly necessary) conventional element. Early on in the dialogue, conventionalism is presented and rejected. Hermogenes espouses a form of conventionalism at the beginning of the dialogue, 33 but Socrates convinces him that names have natural correctness. 34 There has been a troubling trend in the literature to attribute a radical and untenable conventionalism to Hermogenes. 35 The evidence supporting this reading of Hermogenes espousal of conventionalism are his claims that a thing s name is simply whatever it gets called (384c4 385a3), that things may have public and private names (385a4-6), 36 and that a thing s name is whatever each person claims it is and it has as many names as anyone says it has (385d3-e2). As Sedley (2003) points out, it is the endorsement of private (i.e., fluctuating vocabularies) that seems to have painted Hermogenes into a corner. 37 However, Sedley argues that, far from being the voice of extreme conventionalism, 33 The exact scope of Hermogenes conventionalist claims is the subject of some debate. Some authors attribute a radical conventionalism to Hermogenes while other authors attribute a more nuanced view to Hermogenes. This is discussed in section 1.4 of this thesis a-391b is an extended attack on conventionalism. 35 Among the authors that ascribe this view to Hermogenes are Baxter (1992), Palmer (1989), Kahn (1973), and Williams (1982). 36 This discussed on page 4 of this dissertation. 37 Sedley (2003), 52.

26 15 Hermogenes presents a common sense view regarding naming. According to Sedley, the fact that I can make a private decision to assign the name horses to humans does not necessarily undermine linguistic agreement. 38 This interpretation of Hermogenes view is justified on the basis of his anthropological observation that different communities have different names for the same thing (385d8) and his noting the common Greek practice of renaming slaves (384d5). Hermogenes is simply pointing out a fact of the matter that is evident to any person who has engaged in nicknaming. 39 It is clear that every individual or group can assign whatever name it chooses. So, on Sedley s reading of Hermogenean conventionalism private names (i.e. those assigned by an individual or a linguistic sub-community) are as legitimate as public names (i.e. those assigned by the larger linguistic community). I choose to adopt Sedley s interpretation because it not only gives Hermogenes credit for a more interesting and defensible position than is usually attributed to him, but it also makes it possible to connect Hermogenes view with Socrates later statement at 435a7 that one can make an agreement to use a word in a certain way (even if that agreement is with one s self). 38 Op. Cit. p It should be noted that Sedley does indicate that such a practice may stretch the concept of convention. 39 I can, for instance, name my life partner - the most beautiful and perfect woman - Agaphto and use that name in place of the one used to refer to her by my linguistic community. I can use this name for her without undermining the conventions of my linguistic community. Similarly, my older sister can name me Rock-head and use that name to refer to me instead of the name used by my linguistic community (i.e. Rodmon ).

27 16 It may be questioned why Socrates rejects Hermogenes view if it represents a common sense approach to naming. While a view such as Hermogenes could be used to assign names and refer to things, this would not satisfy Plato. The proper account of the correctness of names will require that names do more than merely designate. 40 Consequently, Hermogenes assertion that No name belongs to a particular thing by nature, but only because of the rules and usage of those who establish the usage and call it by that name (384d4-6) will be tested by Socrates and rejected. The rejection of conventionalism as the sole criterion of the correctness of names is tied to the rejection of Protagorean relativism. 41 It should be pointed out that Hermogenes position does not necessarily entail Protagorean relativism. While it is the case that Protagorean relativism entails relativism with respect to naming, one could be relativistic concerning naming while being an objectivist about other truths as Sedley (2003) points out. 42 However, the acceptance of an objective conception of truth forces Hermogenes to relinquish his view that the correctness of names lies solely in conventionalism. 43 Rejecting 40 See section 1.4 of this thesis for full discussion of this claim. 41 Protagoras of Abdera (5 th Century) sophist claimed in the opening line of his work Truth Man is the measure of all things, of the things that they are, and of the things that are not that they are not. While all of the direct sources we have agree on the general subjectivist content of this claim, there is great disagreement as to the degree of subjectivism endorsed here. For excellent discussions of the scope of Protagoras claim see Guthrie (1969), Kirk and Raven (1962), and Barnes (1982). 42 Op. Cit., Modrak (2001), 17 and Sedley (2003), 52.

28 17 Protagorean relativism entails that things have their own objective intrinsic natures. It then follows that there are naturally correct and incorrect ways of acting with respect to them. Since naming is an action and there are naturally correct/incorrect ways to perform any action, I can infer that there are naturally correct and incorrect ways of naming things. It becomes readily apparent then that Hermogenes view is limited to naming having only the force of dubbing or labeling 44 or designation. It is a consequence of Hermogenes view that public and private names it deems legitimate have no necessary connection to any objective truth. Although Hermogenes claims at the beginning of the dialogue are ultimately rejected, 45 they do prepare us for the reemergence of a form of conventionalism in the conversation with Cratylus. Conventionalism reappears in the second half of the dialogue after Cratylus agrees, the success of a word in imitating its object implies, and depends upon, the success of its ultimate components, individual letters, in imitating their respective objects. 46 There are two major problems with this claim. First, there is the issue of how one is to understand the mimetic abilities of names. This question is be addressed in section The second problem, as Socrates points out, is that some words have inappropriate 44 Palmer (1989), A full discussion of the rejection of Protagoreanism in the Cratylus is beyond the scope of this dissertation. For the sake of the discussion in this chapter I provide the following is an outline of the reasoning. 46 Sedley (2003), This is what he calls the Principle of Grounding. Also see Sedley (2003), and Williams (1982),

29 18 elements. 47 Here I focus on the second problem. What Plato wants to know is how we are able to understand each other when we use such words. The word σκληρότης (hardness) is used as an example. Here λ is the problem element. Λ represents softness and is clearly inappropriate. The response given by Cratylus is that usage (ἔθος) allows us to understand each other when I use a word with inappropriate elements. Socrates concludes, both convention and usage contribute something to expressing what we mean when we speak (435b5-6). 48 What, then, is to be made of this exchange? It appears that a resurgent conventionalism won the day. Annas (1982) believes that this concession refutes the naturalist theory held by Cratylus. This view is supported by Plato s statement at 425b1 that expressing something isn t a matter of likeness but of usage, since usage, it seems, enables both like and unlike names to express things. 49 But Sedley (2003) is correct to point out that there are two possible ways to understand 435a-d1: 47 This is first discussed in the conversation with Hermogenes (393d-e). 48 The σκληρότης example is presented within the context of the conversation with Cratylus. 49 On the received reading his is a surprising claim given Socrates dismissal of Hermogenes claim that the correctness of names is by convention. However, I believe that this claim does not commit Socrates to the linguistic relativism espoused by Hermogenes. Socrates is saying that convention or usage must be a part of any true account of the correctness of names. This is different than claiming that the correctness of names lies solely in convention.

30 19 (a) For any name, its success as a tool for communication depends, at least largely, on its imitative powers, but may also depend on a degree of convention; (b) For most names, their success as tools for communication depends on their imitative powers, but there may be some names whose communicative powers depend purely on convention. 50 Sedley argues that Annas among others hold that the σκληρότης case is meant to show that (b) is the case. However, Sedley (2003) points out that the appropriate and inappropriate elements appear in equal number in σκληρότης, and it is use or convention that allows us to break this deadlock. 51 This by no means shows that the imitative powers of names have no part in the correctness of names. It is a combination of the imitative elements in σκληρότης and convention that allows us to understand each other when that word is uttered. Therefore, (a) is the case. The σκληρότης case presents a middle course between conventionalism and naturalism. It contains elements of both views to determine the correctness of names. Over the course of the dialogue, I have seen Plato reject conventionalism as the sole criterion for the correctness of names only to bring a form of 50 Sedley (2003), Op. Cit.,

31 20 conventionalism back to bolster a naturalistic account. 52 It follows from this view that there are degrees of correctness. This is explicitly stated at 435c7 d1. Here Socrates claims that language is best when its sounds are (as much as possible) likeness of things and worst when the opposite is true. 53 The upper bounds of this scale clearly contain the optimal cases in which there are few or none of the sounds are inappropriate. Conversely, the lower bounds of this scale contain the cases where few of the sounds are appropriate. This is supported by an earlier exchange at 432e-433a where an analogy to painting is considered. At the end of this exchange Socrates claims that a name is a name as long as an existing thing s pattern (τύπος) is included (432e), but a name is named well, if it includes all the appropriate letters, and badly, if it includes few of them (433a3-5). It is at the lower bounds of this scale that the role of convention is most prominent in linking a name to an object. As we move up the scale and more of the sounds are appropriate 52 For a very different readings see: Palmer (1989), and Sedley (2003), Palmer argues that Plato endorses a conventionalist theory regarding the correctness of names, because Socrates dissociates himself from it at 426b 428a, and 428d. Also Palmer claims that there is no indication from Plato that certain letters and syllables are to be preferred over other in the formation of names (35). While Palmer is correct that to point out the passages where Socrates dissociates himself from the Cratylan view, it does not follow that Plato has rejected likeness as part of the criteria for the correctness of names. Socrates endorsement of likeness at 435c2-3 is sufficient evidence against Palmer s claims c9.

32 21 likeness of the things named the role of convention diminishes. These claims are examined in detail in the next section. Before moving on it is worth noting that, in subsequent chapters, it will be shown that Aristotle charts a similar course between naturalism and conventionalism. In the opening lines of Peri Hermeneias 1, Aristotle claims the relation between a written mark/spoken sound and an affection of the soul is conventional, but the relation between the affection of the soul and an existing thing is natural. Aristotle s discussion of names relies heavily on naturalistic and conventional elements in way similar to Plato.

33 Ontology and Language Introduction I can now examine (Γ): The correct account of names will explicate the connection between ontology and language. Two major claims are made in the Cratylus regarding the connection between language and reality on which I want to focus: (1) It is claimed that names are tools for dividing being; 54 (2) It is claimed that names are likenesses of the things that are. 55 In section 1.4.1, I discuss claim (1). I focus primarily on the Tool Analogy in the conversation with Hermogenes (386e 390e). In section 1.4.2, I focus on the mimetic theory developed in the later stages of the conversation with Hermogenes and then fleshed out in the conversation with Cratylus The Tool Analogy In the passages that precede the tool analogy, Plato rejects both the Protagorean and Euthydemian theories. 56 Socrates concludes that, it is clear that c e.

34 23 things have some fixed being or essence of their own (386d9-e1). 57 The Tool analogy attempts to show that this conclusion can be extended from πράγµατα (things that are) to praxis (action) done in relation to them. The Tool Analogy is divided into two parts: (a) 386e6-387c5 and (b) 387c6-390e2. 58 In part (a), Socrates 56 The rejection of Protagoreanism (See footnote 39) takes place at 385e4 386d4. Protagoreanism is rejected because such a view would make it impossible, if every individual is the measure of heir own truth and no one can be wrong about anything, to objectively attribute virtue or vice to an individual (e.g., no individual could be said to be truly good or bad). By rejecting these consequences, the Man is Measure Doctrine, which is their source, must likewise be rejected. The rejection of Euthydemian thought, in which everything has every attribute simultaneously, occurs soon after at 386d5 e6. Euthydemianism is rejected this doctrine makes it impossible to attribute virtue or vice to any individual, because this would result in no individual being capable of having either of these attributes without having the other. Again rejecting this consequence of the Euthydemian view entails the rejection of the view that is its source. 57 This on first blush seems similar to or entailed by Hermogenes statement of Cratylan position at the beginning of the dialogue. However, as Baxter (1992) point out, there is an important difference between the Cratylan and Platonic views. Plato emphasizes the objective reality of the thing named (nominata) while the Cratylan position stresses the objective correctness of names. It is because there are stable things with their own being in the world that our naming can serve any useful purpose at all. (39) This marks an important distinction between Plato s position and that of Cratylus and Hermogenes. For Plato ontology comes first. 58 This of course results from following Schofield in transferring 385b1-d1 to follow 387c5. Also see footnotes 3 and 7 of this thesis.

35 24 presents a tripartite schema for the successful completion of an action. 59 The following considerations must be satisfied in order to successfully complete an action: i) The nature of the action ii) The nature of the tool with which the action is to be performed iii) The natural way for that thing (i.e., the thing that is the object of the action) to be acted upon. In considering (i) we must know what our aim is. If our aim is to cut 60 then we need to know what type of cut we wish to make (i.e., perpendicular or horizontal). Also (ii) we need to recognize that successful completion of an action will require the use of the natural tool(s) for that action. One cannot successfully cut with a hammer or hammer with a spoon. Lastly, (iii) the nature of the thing to be acted upon must be considered. The lumberjack is successful in cutting timber because the way she cuts timber is the same way in which it is natural for timber to be cut. It is the nature of the timber that determines this. 61 Similarly with other actions, the nature of the 59 For further discussion in the literature see Baxter (1992), 38-43, Palmer (1989), 83-97, and Sedley (2003), Cutting and burning are two actions that Plato uses as examples at 387a-b. 61 Here I follow Sedley (2003), Sedley points out that the way that it is natural to act on something is identical to the way it is natural for that thing to be acted upon. Sedley draws support for this claim from the Gorgias. (476b-d).

36 25 object must be considered. There are natural ways for each object to be acted upon and these ways are determined by the nature of the object itself. This first section closes with these claims being extended to the act of speaking/naming (387b8-c5). Since speaking/naming is an act like other acts, it too must be done according to the same tripartite schema if it is to be successful. The relationship must obtain among the action (i.e. naming), the accomplishment of the aim of the action (i.e. teaching one another and separating reality), the real nature of the entities named. 62 It follows that we cannot, as Hermogenes claims, name/speak as we wish if our goal is successful speech/naming. The Hermogenean view endorses naming in accordance with our own will and opinions. Any name given by an individual is as legitimate as any name given by any other individual or community. 63 Plato explicitly rejects this at 387b2, 387c1-3, and 387d2-5. Successful naming follows the way and the instrument which the nature of things prescribes (387d5). Only when naming gives proper consideration to the real nature of the thing being named will the naming be successful. Just as it is in other actions, ignoring these considerations leads to failure (387d6). The second part of the tool analogy (387c6-390e2) comes after a short argument for the true/falsity of names (385b2ff). 64 The focus here is on names as 62 Palmer (1989), See section 1.3 of this thesis. 64 See section 1.2 of this thesis.

37 26 tools. Each action has proper tool with which it is successfully accomplished. Plato uses weaving and drilling as examples: Table 1: Tool Analogy Activity Weaving Tool Shuttle Drilling Drill Similarly the name is the proper tool for naming. Next, Plato claims that just as the shuttle is used in weaving to divide the warp 65 and woof 66 that are mixed together (388b), a name is used to separate being from being. There are three possible ways that names can be understood to accomplish this: designation, taxonomy, and analysis. 67 Robinson (1969) argues that names only perform designation. They merely pick out or refer to some object. However, this would not require the skill of the dialectician to determine if names are well made. 68 It would appear that any 65 The threads running lengthwise in the loom and are crossed by the woof in weaving. 66 These are the threads that cross the warp from shelvage to shelvage in weaving. 67 This schema comes from Sedley (2003), However, The taxonomic function originally comes from Kretzmann (1971), Here I refer to Cratylus 388c-e.

38 27 name would be able to perform this function and I would be able to name as Hermogenes claims. The dialectician is the skilled instructor who teaches by separating the οὐσία of things. This requires the other functions listed above (i.e., taxonomy and analysis). Through question and answer (i.e., dialectic), students of the dialectician learn how to divide being properly. This taxonomy is based upon the descriptive content of names. 69 The analytic function of names focuses on a name s ability to describe it nominatum. Names encrypt an object s definitional and ontological components. 70 Thus, an examination of a name will reveal not only how the nature of the nominatum is to be distinguished from other things but also how the internal descriptive content defines what it is to be that thing. Within this second section of the Tool Analogy (387c6-390e2), there is an important set of claims regarding the production of names. Socrates points out that the instructor uses the product of the rule-setter and it is the dialectician who must supervise the work of the rule-setter (388c 390c). Importantly, he claims that the rule-setter is rare among men (388e7ff). This rejects the picture presented by Hermogenes in which everyone is qualified to name things. This establishes that there is a techne (τέχνη) of naming (i.e., attaching the proper names to things according to their essences). The rule-setter uses this techne (τέχνη) to embody in sounds and syllables the name naturally suited to each thing (389d3-4). In doing so, the rule-setter must look to what a name itself is (the Form of name) (389d8). 69 Following Fine (1979). 70 Sedley (2003),

39 28 71 Although the rule-setters in different countries will give each thing the form of the name suited to it (390a4) by using different syllables and sounds, two different names will be equivalent as long as both names embody the same form (389d8-390a7). This is a particularly significant claim. Not only does it surprisingly cut against the Greeks contempt for other cultures in ways similar to ones I encounter in the early parts of Peri Hermeneias, but it also points to a major theme in the Cratylus. In order to name correctly one must look to reality first and then assign names. Ultimately, success in naming depends upon the namers insight into essence of the nominatum and ability to translate insight into letters and syllables (i.e., the τέχνη that allows one to grasp the Form of name). It is worth noting that this is not a descriptive account of language but a prescriptive ideal. Plato s 71 The consensus is that Plato is referring to forms here, but there has been debate in the literature as to the proper was to understand what is being claimed here with respect to the status and function of these forms. There are a variety of views in the literature. Luce (1965) compares this to Republic 596b-597d (the Idea of Bed) and argues the Cratylus position represents an earlier and less complete view of Forms. Calvert (1970) indicates that Plato is distinguishing between the fully transcendent Form Name and the Proper Form Name. Ketchum (1979) argues that the Proper Forms of Name are the features (meaning) of a word by which it a word is true of what is it true of. Anagnostopoulos (1971/2) rejects any connection to transcendent Forms and claims that Plato is referring to the necessary and sufficient conditions which anything must satisfy to be a name. Heath (1888) compares Plato s use of αυτο ὅ εστιν ὄνοµα to Aristotle s τι εν ειναι. Sedley (2003) takes the majority view and takes Plato to be referring to Forms here based on 389b3, 389b10, 390a7, and 390b1.

40 29 objective here is not to describe actual language practices as much as it is to present an account of how language ought to function. By the end of the dialogue, the ability of names to divide reality and provide insight into things is called into question. Words themselves appear to be in conflict with each other. Some words seem to indicate that things are flowing and constantly moving while other names indicate that things are at rest. Socrates points out that we cannot look to names to settle this dispute. It becomes clear that we need to appeal to an extra-linguistic basis to decide this issue. Therefore, we must look to something other than names to discover ontological truths. This, however, is not to suggest the work done here is not valuable. Aristotle, in Peri Hermeneias, explicitly rejects the claim that names are natural tools. 72 Yet, he preserves parts of the tripartite schema Plato offers in the Tool Analogy. Also, I illustrate that other elements of Plato s thought from the Cratylus appear in Aristotle s discussion in Peri Hermeneias. 72 Peri Hermeneias 16a27.

41 The Mimetic/Onomatopoetic Theory The Mimetic/Onomatopoetic theory is the counterpart to the conventionalist theory explored and refuted early on in the dialogue. The majority of the Cratylus is an investigation into and finally a refutation of this theory. The central tenets of the theory are: A) The correct account of names will indicate the οὐσία of each thing (422d1ff) B) Names are vocal imitations (likenesses) of the οὐσία of an existing thing by means of letters and syllables (423e) This account arises in the final stages of the long etymological section of the dialogue (391a-428a). 73 The etymological method of investigation attempts to find the correctness of names in the regularities of language. Under this method the correctness of a name is justified by showing that the name in question is derived from others of similar form which pick out (or signify) similar entities. For example, Poseidon bears his name because "the force of the waves stopped him from walking 73 Much of this discussion is an extended and highly developed etymological exercise. While this material does have philosophical import, a good deal of it is beyond the scope of this dissertation. Consequently, I will only focus on the material that is germane to my discussion. For an in-depth discussion of the etymological aspects of the Cratylus, see Brambaugh (1957), Palmer (1989), and Sedley (2003 and 1998).

42 31 and prevented him from going any further, just like shackles around his feet (δέσµος τῶν ποδῶν)" (402e). Also, Athena is so named because she is the very mind of god (θεοῦ νόησις), or because she has unparalleled knowledge of divine things (Τὰ θεῖα νοοῦσα), or because she has an understanding character (ἐν ἤθει νόησις) (407a7 407c2). Similarly, 'φρόνησις (wisdom) "is the perception (nόησις) of motion (φόρας) and flowing (ῥοῦ)" (411d4). Also, ἑπιστήµη (knowledge) indicates that a worthwhile soul follows (ἕπεται) the movement of things, neither falling behind nor running on ahead (412a1-3). The major problem with this method is the looming danger of an infinite regress. Once a targeted name has been justified via this method the same difficulties (i.e., questions of justification) arise for the antecedents of the targeted name (423d6-e2). To resolve this issue primary names are posited (422b2-3). These primary names are the components (or elements) 74 from which all names are derived. If they are to successfully resolve the issue of infinite regress then primary names must be validated in a way that is different from the one used to validate the names derived from them. It is claimed that derivative names accomplish the indication of the nature of a thing by means of earlier ones (i.e., primary names) (422d8-9). The upshot of this theory is that primary names (letters and syllables) stand in the same likeness relation to a particular essence as the derivative name made of them does. If a derivative name reveals the essence of softness through imitation then the primary names, which constitute that derivative names do so also. In other words, primary names 74 στοιχεῖα.

43 32 accomplish principle (A) (i.e., the indication the οὐσία of each thing) through principle (B) (i.e., the imitation of the nature of a thing) (423a2-3). As can be inferred from principles (A) and (B) above, the validation of primary names is natural. Since primary names play an essential role in this account, I should focus on how they imitate the nature of things. This imitation is first explained through an analogy to gesturing. 75 Vocal imitation is then distinguished from other forms of imitation in two important ways. First, vocal imitation is distinguished from other forms of imitation (e.g., people imitating animals) in that other forms of imitation do not aim at naming the thing imitated. Secondly, vocal imitations are distinguished from music and painting in that vocal imitation is conducted differently than musical imitation and vocal imitation focuses on different objects than musical imitation or painting. Whereas painting and music focus on the graphic or aural representation/depiction of things like color, sound, and shape of entities, naming is not concerned with these phenomena. Naming differs from other forms of imitation in that primary names "imitate in letters and syllables [the] being or essence that each thing has" (423e4-5). This is successfully accomplished when a namer grasps (ἐπιλαµβάνεται) the reality of things that she names by assembling the elements (i.e., letters and syllables). 75 At 423a1ff, the example of using hand gestures to express something heavy or light is presented as is the example of uses gesture to imitate a horse or another animal.

44 33 It may be useful to outline the basic methodology here. Derivative names can be reduced to primary names and primary names are reducible to elementary components (letters and syllables). These basic elements reflect certain properties, qualities, etc. in a self-evident way. 76 For example, rho indicates motion (426d), iota is used to imitate all the small things that can penetrate easily (427a), and lambda indicates gliding (427b). At this point a major question arises: What kind of likeness (imitations) are names? Two positions, which give importantly different answers to this question, are presented in the dialogue. The infallibilist position held by Cratylus claims that: a) Names are imitations of the οὐσία of each thing named b) Names are distinct from the things named c) There is a special relationship between names and the things named such that each name is an exact likeness of its nominatum Cratylus assent to (a) is evidenced by his vigorous agreement, at 430b2, to the claim that names are imitations of things. Also, at 428e1, his support of Socrates statement that, the correctness of a name consists in displaying the nature of the thing it names. 77 Cratylus clearly assents to (b) at 430a4. His endorsement of (c) is 76 As Palmer (1989) p33-34 points out, primary names are perceptibly like the phenomena they signify. They reflect events, properties, qualities, etc. See Sanford (1973) for a discussion of the historical source of this kind of naturalism. 77 Beginning at 427d-428d, the dialogue transitions from a conversation between Hermogenes and Socrates regarding the correctness of names to a conversation

45 34 clear from his assent to the claim that names are given correctly or not at all (429b1-5). 78 Cratylus claims that names are like paintings in that they are imitations of things (430b3), but unlike paintings, in that names are perfect copies of things (430e1). 79 Paintings may include some inaccurate or inappropriate elements, but names cannot include any of these (430e1). Any alteration of a name (e.g. inclusion of inappropriate elements, etc.) makes that name no longer the name of its nominatum (431e9-432a4). One of the benefits of Cratylus view is that it makes names incorrigible. They inerrantly represent the essence of the thing named. Also, this view eliminates the possibility of speaking falsely. 80 On the other hand a fallibilist position, as seen in Socrates response, holds: d) Names are imitations of the οὐσία of each thing named e) Names are distinct from the things named f) Similar to all likenesses (imitations), names fall short of between Cratylus and Socrates on the same subject. The statement at 428e1 marks the beginning of Socrates examination of Cratylus theory. 78 Here Cratylus affirms that names are not given better or worse and that all names have been given correctly. This leads Cratylus to return to Socrates joke at 383c1-3 that Hermogenes is not Hermogenes because he fails to make money. At 429c1-2, Cratylus states that Hermogenes name cannot be Hermogenes, because it is really the name of someone else, namely, the very one who also has the nature. 79 For more on the analogy to painting see Sedley (2000), 45-6, 127-9, and Also, see Baxter (1992), and Palmer (1989), See Robinson (1969), Baxter (1992), and Fine (1977). Also see section 1.3 of this dissertation.

46 35 perfection g) A name is correct if it presents the outline (τύπος) of the thing named h) There are degrees of correctness for names As I have noted earlier Socrates affirms (d) at 428e1. 81 Socrates clearly states (e) at 430a4. Evidence for (f) can be found in the analogy between painting and names (430b6-431e8). The outcome of the examination of this analogy is Socrates claim that there can be better (or worse) name and namegivers just as there can be better (or worse) paintings and painters (431c4-e8). Good paintings (i.e. the work of good painters) includes all the appropriate colors and shapes (431c4) while a poor painting will leave some out or include too many (431c5). If this is true then it follows that some names will not be perfect likenesses. This not only supports tenet (f) but also tenet (h) in that some names will be more correct than others. In response, Cratylus claims that unlike paintings, names cannot withstand any subtraction or addition (431e9-432a4). This sets the stage for the The Two Cratyluses argument for the conclusion that names cannot be perfect likenesses. Tenets (f) (h) are present in the Two Cratyluses argument (432a8-433b7) or are entailed by this argument. Socrates objects that Cratylus treats names as if they were numbers (i.e., any addition or subtraction makes them a different name) (432a5-81 See page 34 of this thesis.

47 36 b1). 82 This, however, has the consequence of collapsing the distinction between names and the thing named (432b2). If names are perfect likenesses then there will be not one thing but two (432b3) and it would be impossible to distinguish the likeness from the thing itself (432d4-7). This is clearly absurd (432d4). For example, there would be two Cratyluses and not one, if a god were to create a perfect likeness of Cratylus in every detail (432b5-c3). Tenets (f) and (h) follow from this; since names cannot be perfect likenesses, there must be degrees of correctness corresponding to the degrees of similarity between the name and the thing named (432e1-5). The important requirement of a name is not perfect likeness but that the pattern of the thing s essence (τύπος) 83 is made manifest (432e7) The example cited by Socrates is the number ten (432a5). If anything is added or subtracted from the number ten then it will no longer be the number ten but instead will be another number. Socrates claims that this account does not hold for things with sensory qualities (432b1). 83 There is a controversy in the literature concerning the proper interpretation of Plato s use of τύπος. Fine (1977), 298 claims that τύπος should be understood as outline of the essence. On this view the τύπος of a thing is an outline or accurate description of that thing s essence. The focus in Fine s view is on the descriptive content of names. What is necessary for a name to be correct is that enough of its descriptive content presents an outline of a particular thing s essence. Palmer (1989) p argues against Fine s reading of τύπος. He takes Fine s view as being committed to the τύπος of a thing being some set of descriptive properties. Palmer claims that while other instances of τύπος in the Cratylus are translated as outline, it is not the case that these instances refer to the descriptive content of names. At 397a3, Socrates claims that he and Hermogenes have discovered a τύπος

48 37 By the end of the conversation with Cratylus, the infallibilist position is shown to be untenable. The fallibilist position wins the day and the stage is set for the re-emergence of conventionalism. As discussed in section 1.3, convention and usage play a role in our ability to understand names that contain disparate elements. 85 I demonstrate in chapters 3 and 4 that Aristotle develops his own account of how ὄνοµατα are likenesses. The exact nature of this account has been the subject of controversy in the literature, but I illustrate that it is indebted to Plato s work in the Cratylus. There is one more topic that needs to be explored regarding the connection between naming and ontology in the Cratylus. In the later stages of the dialogue the (i.e. outline) for their subsequent investigation. The use of τύπος here does not point to the descriptive content of names. Palmer argues that the use of τύπος at 436e7 should be understood as a non-technical synonym for οὐσία (116). Palmer cites Socrates claims at 394a3-b6 and 433a4-6 as evidence for his view. Here Palmer claims that it is a names ability to present the οὐσία of a thing that makes a name correct not its descriptive content. What both of these interpretations have in common is that the τύπος of a thing being intimately connected to that thing s essence. I side with Palmer because it appears that Plato has in mind here something more than merely some set of descriptive properties. This is supported by Socrates claim at 436e2-4 that names signify the being or essence of things In addition, Palmer s interpretation is consistent with the use of τύπος by Plato in other works (e.g., Republic 396e9 and 443c1). 84 This claim is reiterated at 433a See section 1.3 of this thesis.

49 38 claim made that names are tools for instruction and dividing being is examined. 86 Cratylus claims that names are the best and only way of giving instruction about the things that are (435e-436a). This claim by Cratylus is not surprising given his previous claims and the discussion that immediately preceded this. This would seem to follow from the Tool Analogy and the points just discussed regarding the mimetic properties of names. However, Socrates disputes this (436b-d). He points out that the giver of the first names could have used an erroneous ontology to develop names (436b). Cratylus protests that the consistency of names testifies to the truthfulness of the ontology used by the first name giver (436c). Socrates rejects Cratylus claim by pointing out that consistency does not preclude error; if one starts from a point of error then the error will persist even though one s thought may be consistent (e.g., geometrical constructions) (436d). Socrates also questions whether names are truly consistent with one another. If we assume that all things are in motion, then there are some names that signify that existing things are at rest (437a-c). For example, 'ἱστορία' ('inquiry') signifies "the stopping (ἵστησι) of the flow (ῥοῦν)" (437b1-2). Cratylus responds that most of the names indicate motion, but Socrates rejects this idea because names should not be tallied like votes to determine correctness (437d). Sedley (2000) points out that it is easy to make the mistaken inference here that the entire etymological procedure should be rejected This claim is first made at 388c1 during the conversation with Hermogenes. Also, see section of this thesis. 87 Sedley (2003),

50 39 However, as Sedley correctly points out, Socrates explicitly endorses the exegetical correctness of etymologies. 88 Sedley is careful to point that while Socrates believes that the etymological method 89 accurately captures the beliefs of the name givers, it is not philosophically adequate. 90 [T]he contradiction which Socrates claims to have revealed is not an exegetical one but a philosophical one. 91 The ability of the etymological method to teach us philosophical truths is called into question. This contradiction shows that the name givers had insufficient understanding. 92 Consequently, Socrates suggests that one must look to something other than names (438d2-e1). Cratylus agrees that this is the correct course of action. However, Socrates points out that if this is the case and it seems that we are able to learn ontological truths from a source other than names then it would appear to be better to learn ontological truths directly than through examining likenesses (i.e., names) (438e4-b1). I argue in the chapters that follow that Aristotle s discussion of language in Peri Hermeneias relies on the ontological schema developed in the Categories. The work done in the Categories allows for Aristotle to make a clear connection between language and ontology. 88 Ibid. 89 See footnote 73 on p31 of this thesis. 90 Ibid. 91 Sedley (2003), Ibid.

51 40 Summary and Findings I have shown that Plato, in the Cratylus, rejects both conventionalism and naturalism as the sole criteria of the correctness of names. Conventionalism is rejected because it fails to discharge important functions of naming. While conventionalism can provide a way to refer to things (i.e., the function of designation), it does not allow the namer to perform other important functions of naming such as analysis and taxonomy. 93 As I have demonstrated, Plato also rejects naturalism because there is no non-arbitrary way to decide (on a linguistic basis) which ontological framework names should represent. This presents us with a necessary pre-condition for any putative account of the correctness of names: that is, it provides an objective ontological and epistemic taxonomy (explanatory framework) of existing things. This follows from not only the rejection respectively of conventionalism and naturalism as sole criterion for the correctness of names, but also from the fact that correct names communicate the nature of the nominatum. Because they signify the οὐσία of the thing named, names are able to divide reality by picking out the basic units of reality or by identifying instantiations of these units. Hence, the importance of an objective ontological framework is established. Also, I have explained Plato s rejection of Cratylus claim that names are exact copies or likenesses (ὁµοίωµατα) of things. This rejection has implication for and connections to Aristotle s claims in Peri Hermeneias. I need not adopt the controversial thesis that Aristotle had the Cratylus in focus when he wrote Peri 93 Sedley (2003), Also see section of this thesis.

52 41 Hermeneias in order to elements in Aristotle s account that have similarities to the final Socratic position. In the chapters that follow I argue that Aristotle is committed to a conception of non-perfect mimesis concerning mental images. This allows him to avoid problems presented by the Two-Cratyluses Argument.

53 42 CHAPTER TWO - The Greek Commentary Tradition and The Themes of Peri Hermeneias 1-6 Introduction Having completed our examination of Plato's influential work in the Cratylus, I am now in a position to embark on an exploration of how Aristotle addresses the key questions left unanswered by Plato in the Cratylus. The discussion focuses on Aristotle s semantic thought in Peri Hermeneias 1-6 because it contains Aristotle s only explicit attempt to define meaning. 94 In order to illuminate the important themes and interpretive challenges of these passages, I begin by examining key parts of the tradition of Greek commentaries on Peri Hermeneias. One of the challenges that confronts any attempt to engage the Greek commentary tradition is that one cannot be sure that the commentators were consistently working from the same set of source materials. Over time the text of Peri Hermeneias passed through many hands before reaching the authors of the commentaries. 95 When approaching the text of Peri Hermeneias through the Greek commentary tradition, one must be aware of the ways in which variations in the received text can affect the claims in Peri Hermeneias. Another obstacle is the interpretative decisions made by the commentators. In many instances, commentators consciously or subconsciously conjoin their 94 Modrak (2001). 95 For a full discussion of the transmission of Aristotelian text, especially Peri Hermeneias, see Sorabji (1990 and 2005), Arens (1984), and Edel (1982).

54 43 philosophical commitments with Aristotle s work. Prime examples of this are the prochronistic 96 attribution of later views to Aristotle and the commitment of Neo- Platonic commentators to the harmonization of Plato and Aristotle. In order to highlight and whenever possible avoid these obstacles one needs a translation that neutrally interprets the meaning of ambiguous terms and includes material from textual variants. Therefore, as this chapter proceeds, I present translations of relevant portions of the text of Peri Hermeneias 1-6 with key passages, terms, and points of controversy highlighted. These translations are based mainly on the translation done by J.L. Ackrill, but also include, in the form of footnotes and parenthetical insertions, translations by Deborah Modrak (2000), Harold Cooke (1938), and myself. 97 In the sections that follow I use these translations to critically evaluate the work of the early Greek commentators. I chose the Early Greek commentary period because it is our closest source to the time of Aristotle and therefore gives us important insight into how the claims of Peri Hermeneias were understood at that time. I begin this chapter by briefly outlining the principal themes of Peri Hermeneias 1-6. Next, in section 2.2, I critically analyze attempts by the early Greek commentators to explain the claims of Peri Hermeneias. In this section, I focus mainly on Ammonius commentary since it is the earliest extant commentary on Peri Hermeneias. I also examine the references to the work of the 96 Prochronism is a chronological error in which an event, concept, person, etc. is ascribed to a date earlier than the actual one. 97 I have color coded the parenthetical insertions as follows. green= Modrak, blue = Cooke, and purple = myself.

55 44 early Greek commentators found in Boethius commentary. 98 In section 2.3, I review and summarize the findings of the previous sections and then I set the stage for the chapters that follow Peri Hermeneias 1-6 Outline of the Principal Themes in Peri Hermeneias 1-6 Ch. 1 The relation of language to mental states and things in the world; truth/falsity found in combination Ch. 2 Definition of name; simple v. composite names, indefinite names, Inflections of names Ch. 3 Ch. 4 Ch. 5 Ch. 6 Definition of verb, indefinite verbs, inflections of verbs Definition of a sentence, Definition of a proposition Simple and complex/composite propositions Affirmations and negations 98 Boethius work clearly is not part of the early Greek commentary tradition. His work of course belongs to the tradition of Latin commentaries. For our purposes in this chapter, we will only focus on his citation of the work of early Greek commentators.

56 The Early Greek Commentary Tradition Much of the work produced by the early Greek commentators has been lost. There is evidence that Aspasius, Herminus, Alexander of Aphrodisias, Porphyry, Iamblichus, and Syrianus produced commentaries on Peri Hermeneias that are now lost. 99 However, two later commentators present us with a valuable resource on the early Greek commentary period. Ammonius commentary is not only the earliest extant commentary but it also draws material from and engages the work of earlier authors. As such Ammonius commentary on Peri Hermeneias gives us key insights into the way in which the claims made by Aristotle were understood and shows us interpretative controversies that arose. Another source of material on this period is Boethuis s commentary 100 because it also discusses the work of earlier commentators and provides us valuable insight into their thought. In 99 Both Arens (1984) and Sorabji (1990 and 2005) have excellent discussions of the early Greek commentators and their works which are no lost to us. 100 Boethius s commentary belongs to the Latin commentary tradition. It is beyond the scope of this dissertation to explore this tradition. Instead I will examine Boethius s reports of the views of the early Greek commentators.

57 46 that follows, I examine excerpts from these commentaries and then discuss general themes found in them. 101 As I indicate in chapter three, recent commentators have noted the difficulty of properly interpreting the claims made in Peri Hermeneias 1-6. This however is not a new problem. Ammonius begins his commentary by noting the difficulty of this text. 102 Boethius also acknowledges this book here is more than all others a condensation of acute thinking and brief expression. 103 It is not surprising, then, that both Ammonius and Boethius struggle to understand the claims made in Peri Hermeneias and that this problem was also present for the earlier Greek commentators that they cite. 101 A full discussion of Boethuis commentary and Ammonius commentaries is beyond the scope of this dissertation. Therefore, I have limited my examination of these commentaries to the sections that deal with the claims made with respect to Peri Hermeneias Ammonius1, Arens,

58 The Schema of Chapter One of Peri Hermeneias At Peri Hermeneias 16a3ff, Aristotle claims: Now spoken sounds [φωνῇ] (the things that are in the voice/sound) 104 are symbols [σύµβολα] of affections in the soul [ἐν τῃ ψυχῇ] and written marks [γραφόµενα] symbols of spoken (the things that are in the voice/sound) sounds. And just as written marks are not the same for all men, neither are spoken sounds. But what these [ταυτα] 105 are in the first place [πρώτως] signs [σηµεῖα] of - affections of the soul [παθήµατα] - are all the same [ταυτα] for all [πᾶσι]; and what these [ταυτα] affections are likenesses (representations or likenesses, images, copies) [ὁµοίωµατα] of The word φωνῇ is translated by most scholars as spoken sounds or words spoken. This seems to be supported by the contrast with γραφόµενα at 16a5. However, because I believe that this translation masks an important contrast in the text, I have offered an alternative translation. In part, as we will see I am following the earlier Greek commentators in noticing that this phrase set s up the contrast made later between significant sound and the inarticulate noises of animals at 16a28. This will be discussed in more detail in section Magee (1989) points out that there is considerable controversy over what accents are to be used with respect to the occurrences of ταυτα in this section of the text. This will be discussed more in section

59 48 actual things - are also the same [ταυτα]. These matters have been discussed in the work on the soul and do not belong to the present subject. Both Boethius and Ammonius identify three items 106 discussed in the first chapter of Peri Hermeneias that they consider to be important to understanding the claims made by Aristotle 107 : things (πράγµατα), thoughts (πάθηµατα), 108 and vocal sounds (ὄνοµατα and ῥῆµατα). 109 These terms are at the center of early commentators attempts to understand the claims of Peri Hermeneias 1-6. I show in chapter three of this dissertation that some modern commentators have a similar schematization of this part of the text. 110 In this section, I examine the role of πάθηµατα and πράγµατα. Vocal sounds (ὄνοµατα and ῥῆµατα) are discussed in section As I reveal in chapter three, many modern commentators focus a major portion of their investigations on the claims made in Peri Hermeneias 16a3-8. The early Greek commentators also dedicated a good amount of time to discussing this part of the text. Both Ammonius and Boethius give attention to the phrase "the 106 Ammonius adds letters (στοιχεία) to this list that follows. 107 Boethius 20 and 21 and Ammonius 18, Boethius refers notions or concepts. 109 Boethius refers to spoken forms or words. 110 Kretzmann (1974), for one, identifies seven semantic elements/relations in this section of the text.

60 49 things that are in the voice (ἐν τὰ φωνῇ)" at 16a Ammonius questions why Aristotle uses this phrase. 112 He points out that Aristotle said, "What is in vocal sounds are symbols of affections of the soul" instead of saying "vocal sounds are symbols of affections of the soul." 113 At first glance, this may seem to be a minor point, but when carefully examined, the interpretation of this phrase can have a major impact on how the claims of Peri Hermeneias are understood. Ammonius argues that Aristotle uses this phrase to set up the distinction between voice (in general) and significant vocal sounds. Ammonius claims that being a symbol and by imposition in general do not fit the vocal sound simpliciter, but the name and the verb. 114 The ability to produce vocal sounds, just like other abilities (e.g., hearing, seeing, tasting, etc.), is given to us by nature, but names and verbs are the products we, as humans, create (not nature). Just as a gold coin is created when natural material (i.e., gold) is molded and impressed in certain ways, Ammonius claims, names and verbs are not simply vocal sounds, but vocal sounds shaped and formed by linguistic imagination (λεκτική φαντασία) and accepted as symbols (σύµβολα) of the thoughts of the soul. 115 Additionally, the early commentators discussed Aristotle s use of sign (σηµεῖόν), symbol (σύµβολον), and likeness (ὁµοίωµατα). Ammonius claims that 111 Ammonius 22,4ff and Boethius II 42.6ff , 4ff , , , 1-2.

61 50 sign and symbol are synonymous in that both terms stand for conventional tokens. 116 I argue in the next chapter that some modern scholars hold a similar view. It follows from this that at least some of the early commentators did not distinguish sign from symbol. The early commentators did distinguish sign or symbol from likeness. Ammonius states that, Likeness differs from symbol in that it [i.e., likeness] wants to image the very nature of a thing as far as possible and it is not in our power to change it. 117 Ammonius clearly believes likenesses to be natural images of things (πράγµατα) and thereby fundamentally different from symbols. At 16a2-8, Aristotle claims that what is in the voice are σύµβολα of πάθηµατα in the soul. The early commentators struggled to understand the phrase 'πάθηµατα of the soul'. 118 Ammonius reports that Andronicus of Rhodes doubted the authenticity of Peri Hermeneias because there is no reference to affections (πάθηµατα) of the soul in De Anima. 119 Ammonius ultimately rejects Andronicus charge on the grounds that the schema in Peri Hermeneias 16a2-8 is analogous to the account of φαντασία in De Anima. According to Ammonius the πάθηµατα of the soul in Peri Hermeneias 16a3ff are analogous to the φαντάσµατα in De Anima , , Here I follow Wheeler (1995) p 195 and assume that there is no relevant philosophical difference between affections of the soul at 16a6-7 and affections in the soul at 16a , 25 7,14.

62 51 Ammonius points out at De Anima III.7 (431b2) that Aristotle states, the faculty of thinking (τὸ νοητικόν) thinks the forms in the images (φαντάσµατα). 120 These images are formed when the imagination is impressed with the forms of existing things during perception. Boethius presents a different understanding of Andronicus error with respect to the reference to De Anima. Boethius claims that Andronicus questioned the authenticity of Peri Hermeneias because De Anima 432a10-14 has no discussion of passiones animae (πάθηµατα) in the sense of affectiones (affections/dispositions). 121 Following Alexander, Boethius corrects Andronicus by pointing out the passiones are not understood as affectiones but as intellectus (νόηµα). 122 Ammonius also identifies πάθηµατα with passions of the soul or thoughts (νοήµατα). 123 As discussed earlier, Ammonius claims these thoughts are mental images of things engraved in the soul. 124 The view that πάθηµατα of the soul are thoughts (or experiences of the soul) is not unique to Ammonius. Boethius indicates that Aspasius, Herminus, and Porphyry considered πάθηµατα to be experiences of the soul (i.e. thoughts (νοήµατα)). 125 All these views 120 6, Magee (2003), 31 footnote Ibid , and 20, , , and Sorabji (2005), Sorabji points out that the basis for this view is the following passage from Republic 2 (382B9-C1): Since the παθήµα in words is a sort of imitation (mimema) and a subsequently generated image (eikon) of the experience (παθήµα) in the soul, it is not altogether unmixedly false.

63 52 are similar in that they take πάθηµατα to be passive states or something that the soul undergoes. One problem with this view is that these commentators identify πάθηµατα with thoughts, but it is not the case, for Aristotle, that all thoughts are passive. 126 The correct account of πάθηµατα in the soul is an issue that remains unresolved to this day. As I demonstrate in chapter three, some of the competing modern interpretations of 'affections in the soul' in the contemporary literature bear a striking resemblance to the views of these early commentators. The claims made at 16a6-8 presented a particular set of problems for these commentators. The way in which they addressed these problems highlights important features of the text. Some commentators questioned the proper way to understand the claim that πάθηµατα are the same for all (16a7). According to Boethius, Aspasius questioned how πάθηµατα could be the same for all men when there are diverse opinions on issues such as what is good. 127 Aspasius reported solution is to claim that Aristotle did not speak of impressions of the soul concerning immaterial things, but only concerning what could be perceived by the senses. 128 The importance of this move by Aspasius is that it connects Aristotle s claims regarding πάθηµατα of the soul with his epistemology and account of perception. In chapter three I indicate that this insight plays a key role in modern discussion of Aristotle s conception of πάθηµατα. Herminus, Alexander of 126 Ibid. 127 Ibid. 128 Ibid

64 53 Aphrodisias, and Porphyry addressed a deeper and more troubling issue concerning this section of the text. As Magee (1989) points out, the original Aristotelian text most likely lacked accents. 129 Consequently, some of the early Greek commentators were concerned with the ambiguity of the unaccented text and they suggested which accents to apply to the written text. According to Boethius, Alexander argued for the use of ταὐτὰ ( the same ) instead of ταῦτα connection with ὁµοίωµατα at 16a Alexander is reported as claiming that Aristotle s intention here is to distinguish between those things which are significant by convention and not the same for all (i.e. φωναί and γράµµατα) and those things which are natural and the same for all (i.e., ὁµοίωµατα and πράγµατα). 131 However, Alexander s reading commits Aristotle to the view that the same likeness are of the same things. 132 Herminus rejects Alexander s account. 133 Herminus is reported as suggesting ταῦτα instead of ταὐτά at 16a6 and 8 with respect to πάθηµατα and πράγµατα. 134 He claims that this passage is best understood as: Now what these 129 Magee (1989), Boethius 40, 25ff. 131 Ibid. 132 Waitz (1844) Vol. I p and Magee (1989) p9-10. Magee points out that In drawing out the antithesis, so Alexander appears to have held, ὁµοίωµατα, following the expression, ταὐτὰ παθήµατα τῆς (16a6-7) and taken in coordination with πράγµατα ταὐτά (16α7-8), accurately represents Aristotle s meaning only if read ὀξυτόνως ( the same, ταὐτά). 133 Boethius 39, 25-40, Magee (1989), 9-10.

65 54 are in the first place signs of, these are experiences of the soul for all and what these experiences are likenesses of, these too are things. Herminus claims, all we must take from this is that experiences of the soul are what spoken sounds signify, and what has likenesses and is contained in thoughts are things. 135 In response to Aspasius, Herminus argued that the experiences of the soul, which spoken sounds signify, are not the same for all, since some words are equivocal and prompt more than one concept or thought in the minds of those who hear them. Porphyry is reported as challenging Herminus position by claiming that [S]ince Aristotle distinguished conventional from natural spoken sounds on the grounds that the former were not the same for all men, he would have to maintain that what he thought was natural was natural because it was the same for all men. 136 It appears on Porphyry s view that what is the same for all is not the individual πάθηµατα symbolized by (associated with) some conventional vocal sound, but the fact that πάθηµατα are images of πράγµατα. On Porphyry s view the same for all claim is vital to Aristotle s view. In response to Aspasius, Porphyry s solution is when each understands something different from the single name, the one who said the equivocal name designates more clearly what he meant to signify by the name, the listener then receives it and both agree on the one thought, which in turn becomes one for those 135 Boethius 39,25ff (Sorabji p207). Again this is Boethius report of Herminus claims ,9ff.

66 55 who at first had had different experiences of the soul because of the equivocation of the name. 137 The importance of this series of moves is that it points to two key points regarding this text: (1) The lack of accents makes what Aristotle originally wrote ambiguous (2) This ambiguity requires that careful interpretive decisions 138 The first is point clearly demonstrated in interpretative moves above. In particular the moves made by Alexander, Herminus, and Porphyry with respect to the occurrences of ταυτα, show how dramatically different interpretations can arise from interpretative decisions. Another important feature of these moves is that they point out the importance that material from De Anima plays in understanding the claims made in Peri Hermeneias. Before I leave this section it is worthwhile to examine the programmatic statement at the beginning of Peri Hermeneias. At 16a1 Aristotle states: First we must settle what a name [ὄνοµα] is and what a verb [ῥῆµα] is, and then what a negation [ἀπόφασις], an affirmation [κατάφασις], a statement [ἀπόφανσις], and a sentence [λόγος] are. 137 Ibid. 138 By interpretative decisions, I mean not only the choices made by the concerning the application of accents to the Aristotelian text, but also the cluster of decisions made as part of hermeneutics.

67 56 The early Greek commentators focused on this opening statement of Peri Hermeneias. Although this section of the text has mostly been overlooked by some modern commentators, it is clear that some of the early Greek commentators used it to support the attribution of a tripartite schema for φωναί (i.e. writing, speech, and thought) to Aristotle. Boethius reports that Porphyry attributes a similar tripartite schema for speech to Aristotle. He claims, For the Peripatetics there are three [kinds of] speech, one written in letters, a second expressed vocally, the third composed in the mind. 139 Likewise Ammonius claims, since names and verbs are primarily parts of speech, there will be some verbs and names that are written, some spoken, some exercised in the quiet of the mind. 140 Ammonius claims that Aristotle wants to discuss names and verbs first because they and the sentences composed of them are seen in three ways: in the soul according to simple thoughts (νοήµατα) and the so-called internal speech (ἔνδθεν λόγος), in the actual voicing (ἔκφηµι), or in the writing (as we also say of written items that one is a name, another a verb, another a sentence). Thus, since both names and verbs, which he said one must set out, are seen, as we said, in three ways, in being thought, said, or written, for this reason he said [16a3-4] in this way that what are in the vocal sounds are symbols of the thoughts in the soul, which he 139 Boethius 29, 29 30, Ibid.

68 57 says are its experiences for the reason given at the beginning, and that written marks are in turn symbols of what are in the vocal sounds [16a4] Φωναί (Spoken Sounds): Names (ὄνοµατα) and Verbs (ῥῆµατα) In this section, I examine, in turn, the commentators accounts of names (including indefinite names) and verbs (including indefinite verbs) because Aristotle listed them in his programmatic statement at 16a1ff and they have a central role in his discussion of language in Peri Hermeneias 1-6. At 16a Aristotle states, A name (noun) [ὄνοµα] is a spoken sound [φωνὴ] significant by convention [σηµσντικὴ κατὰ συνθήκην], without time [ἄνευ χρόνου], none of whose parts is significant in separation [ἧς µηδὲν µέρος ἐστὶ σηµαντικὸν κεχωρισηένον]. This section of the text also generated a good deal of discussion. I follow Ammonius and begin by examining the claim that ὄνοµατα are significant by convention (σηµαντικὴ κατὰ συνθήκην). Ammonius claims that significant is included by Aristotle to distinguish names from meaningless vocal sounds. 142 On 141 Ammonius 22, ,

69 58 this basis he attributes to Aristotle a four-fold schema for vocal sounds: significant, meaningless, lettered, unlettered. Figure 1 Schema for Vocal Sounds Vocal Sounds 12,$3 Significant *+µ$,%-./' Meaningless (0+µ$,%-./' Lettered!"#µµ$%&' Unlettered ()"#µµ$%&' Lettered!"#µµ$%&' Unlettered ()"#µµ$%&' The species of vocal sounds are significant and lettered, significant and unlettered, meaningless and lettered, meaningless and unlettered. Ammonius gives the following examples of each type of vocal sound. 143 Human vocal sounds are significant and lettered and the sounds made by animals are significant but unlettered. Ammonius cites blituri as an example of a vocal sound that is meaningless and lettered and whistling which is done for no reason as an example of one that is meaningless and unlettered , , 15.

70 59 This schema does seem to have some textual support. Aristotle s statement at 16a28 that the ἀγράµµατοι ψόφοι 145 of beasts reveal something but none of them is a name appears to support Ammonius view. Ammonius also focused on the claims that: 1) names are significant by convention (16a20); 2) no name is a name naturally (16a29). It is here that Ammonius asks how, when Socrates in the Cratylus argues against Hermogenes assertion that names are by imposition and shows that they are by nature, can Aristotle insist that in these words that no name is by nature? 146 After reviewing the positions put forth in the Cratylus, 147 Ammonius claims Plato s views in the Cratylus and Aristotle s views in Peri Hermeneias are similar. According to Ammonius, Plato s view is a compromise between Hermogenes view and Cratylus view in that names are by imposition, because the name giver assigns them, and they are by nature, since some names indicate the universal and simply eternal. Ammonius claims that Aristotle also steers a similar middle course between Hermogenes account and Cratylus account in that he denies of them the sense of by nature which the Heracliteans were advocating, just as Plato did, and he would not have declined to call them by nature in the same sense as the divine Plato does For further discussion see Zirin, R. (1974) , , 22 36, ,

71 60 Ammonius clearly views Aristotle s account as a corollary to the mimetic account found in Plato s Cratylus. 149 This is an example of what Sorabji (2005) notes as a general trend among the Neo-Platonic commentators on Aristotle: the attempted harmonization of Plato and Aristotle. 150 Ammonius account is deficient in two major respects. First, it is not clear, by the end of the Cratylus, whether or not Plato endorses the mimetic account. 151 Secondly, it is not clear that Aristotle would endorse this account. With respect to Aristotle s claim that no part of a name is significant in separation (16a20), Ammonius claims this is used to distinguish names from vocal sounds constructed out of several names (e.g., rational mortal animal ). 152 Unlike vocal sounds constructed out of several names, the parts of individual names do not signify in separation (e.g., the syllable ἄν in the name ἄνθρωπος ) Ammonius clearly states this at 40, Sorabji (2005), See Section of this dissertation ,25ff , 30ff.

72 61 This leads to Ammonius discussion of Aristotle s account of compound names. At 16a21ff Aristotle states: For in 'Whitfield' ( Goodsteed ) [Kάλλιππος] the 'field' ( steed ) [ἵππος] does not signify anything in its own right [ουδὲν αὺτο καθ ἑαυτὸ σηµαίνει], as it does in the phrase 'white field' [καλὸς ἵππος] ( good steed ). Not that it is the same with complex names [ἐν τοῖς συµπεπλεθµένοις] as with simple ones [ἐν τοῖς ἁπλοῖς ὀνόµασιν] 154 : in the latter the part is in no way significant [οὐδαµῶς σηµαντικόν], in the former it has some force (wish/pretend to signify) [βούλεται] but it is not significant [οὐδὲν σηµαίνει] of anything in separation, for example the 'boat' [κέλης] in 'pirate-boat' [ἐπακτποκέλης]. Although Ammonius takes compound names to be somewhat similar to phrases constructed out of names, there is a still a marked difference between the two accounts. The ἵππος part of the name Κάλλιππος does not have the same signification as ἵππος in the phrase καλὸς ἵππος. Ammonius claims the former 154 Cooke (1938) translates this, as It is necessary to notice, however, that simple nouns differ from composite.

73 62 signifies an animal of this kind [i.e. horse], 155 while the latter does not have the same force because it is part of a name that indicates the simple thought of the man whose name this is. 156 When broken off from its proper whole, it becomes a dead thing as far as signifying is concerned, no different from the totally meaningless syllables which are the parts of simple names. 157 However, the difference is that the parts of compound names give an impression of signifying 158 Ammonius interpretation of compound names highlights the connection between thoughts and vocal sounds. On Ammonius view, the signification of a name is at least in part a function of the relationship between thoughts and vocal sounds. Ammonius presents the examples of Neapolis (New City) and Helioupolis (Sun s City). He indicates that when the parts of these names are taken as significant, there will be a sentence signifying either the city now founded or that sacred to the sun, each part indicating its own idea. 159 He goes on to point out that when the parts are no significant by themselves (as is the case with compound names), there will be a name that, as a whole, signifies one thought. 160 The reason that the parts of compound and simple names do not signify in separation is that simple thoughts do not themselves have parts. In the chapters that , , , , , , 6-8.

74 63 follow, I show that modern scholars also have focused considerable attention on Aristotle s account of names in Peri Hermeneias and its possible semantic import. Indefinite names are the next topic, with respect to names, discussed by Aristotle in Peri Hermeneias and examined by Ammonius. Aristotle states at 16a30ff: 'Not-man' [οὐκ ἄνθρωπος] is not a name (and the like are not nouns), nor is there any correct name (established/set down name) [κεῖται ὄνοµα] for it (I know of no recognized names we can give such expressions). It is neither a phrase [λόγος] nor a negation [ἀπόφασις] (which are neither denials nor sentences). Let us call it an indefinite name (indefinite nouns) [ὄνοµα ἀόριστον] (since we use them of all kinds of things, non-existent as well as existing) [καὶ ὄντος καὶ µὴ ὄντος]. Ammonius account of indefinite names is fairly straightforward. He begins by explaining that Aristotle discusses indefinite names, because he saw that such vocal sounds too were often included in assertions. 161 Ammonius states that indefinite names are not names because a name signifies one nature, while each [indefinite name] destroys [ ] what is signified by the name without negative [particle]. 162 Ammonius goes on to point out that indefinite names introduce everything other than the name with the negative particle. Not-man, for example, applies to horses, dogs, goat-stags, and hippocentaurs. These items have in , ,

75 64 common the lack of being what man is. [Indefinite names] signify one thing in a way, namely everything besides the definite thing considered as one. 163 Overall, Ammonius does not view Aristotle s account of indefinite names as a problem. He sees it merely as another part of Aristotle s investigation of statements. I indicate in chapter four that Ammonius account of indefinite names is similar to that of some recent scholars. I argue that while Aristotle account of is much more defensible than some commentators have given him credit. I am now in a position to examine the early Greek commentators views on the next item in the programmatic statement at 16a1-2 (i.e., verbs). At16b5 11, Aristotle claims: A verb [Ῥῆµα] is what additionally signifies time [δέ ἐστι τὸ προσσηµαῖνον χρόνον], 164 no part of it being significant separately (No part by itself has a meaning) [οὗ µέρος οὐδεν σηµαίνει χωρίς]; and it is a sign of things said of something else [καί ἔστιν ἀεὶ τῶν καθ ἐτέρου λεγοµένων σηµεῖον]. It additionally signifies time: 'recovery' ( health ) [ὑγιεια] is a name (noun), but 'recovers' ( is healthy ) [ὑγιαίνει] is a verb, because it additionally signifies something's holding now [προσσηµαίνει γαρ τὸ νῦν ὑπάρχειν]. And it is always a sign (indication) [σηµεῖον] of , Cooke (1938) translates this as a sound which not only conveys a particular meaning but has a time-reference also.

76 65 what holds, that is, holds of a subject (of the things belonging to a subject or in a subject) [οἷον τῶν καθ ὑποκειµέµου ἢ ἐν ὑποκειµένῳ]. Ammonius begins his analysis of this passage by claiming that Aristotle passes over everything the verb has in common with the name for the sake of brevity. 165 Ammonius believes that this explains why Aristotle starts his discussion of verbs by stating one of the key ways in which verbs differ from names (i.e., verbs signify time) 16b7-8. Porphyry claims that the purpose of Aristotle s statement at 16b7-8 is to distinguish verbs from sentences composed of verbs (e.g., To walk is to move ). 166 According to Ammonius, verbs signify (i.e., primarily indicate) actions or passions. 167 The reason that verbs additionally signify time is that the times adjunct to the actions or passions, according to which they are said to be, to have been, or to be going to be, we see along with these actions and passions as appearing alongside them. 168 Next, Ammonius indicates that verbs always occupy the place of the predicate in a proposition because Aristotle indicates that verbs are always a sign of something said of something else. 169 Ammonius connects these claims with Aristotle s work in the Categories, specifically stating that some verbs belong to , , , , ,

77 66 the things of which they are predicated according to their essence and are said to be predicated as completing them, which he [Aristotle] called of a subject while others belong as accidents to their subjects, which we refer to as in a subject. 170 This is an important move by Ammonius. It relates Aristotle s discussion of language in Peri Hermeneias, with its important reference to the material in De Anima, to his discussion of predication and ontology in the Categories. In the next chapter, I argue that the ontological schema and account of predication in the Categories can help us understand the claims made by Aristotle regarding the signification of names, verbs, statements, existing things (πράγµατα). In addition I demonstrate that the connection between the Categories and Peri Hermeneias allows provides Aristotle with the ontological underpinnings that I found lacking for Plato in chapter one , here Ammonius points to Categories 1a20ff. 171 See Section of this thesis.

78 67 Ammonius moves on to an examination of Aristotle s account of indefinite verbs. At 16b11-15 Aristotle states: 'Does not recover' ( Is not-well ) [οὐχ ὑγιαίνει] and 'does not ail' ( is not-ill ) [οὐ κάµνει] I do not call verbs. For though they additionally signify time and always hold of something, yet there is a difference - for which there is no name [προσσηµαίνει µὲν γὰρ χρόνον καὶ ἀεί κατά τινος ὑπάρχει, τῇ δέ διαφορᾷ ὄνοµα οὐ κεῖται]. 172 Let us call them indefinite verbs [ἀλλ ἔστω ἀόριστον ῥῆµα], because they hold indefinitely of anything whether existent or non-existent [ὅτι ὁµοίως ἐφ ὁτουοῦν ὑπάρχει, καὶ ὄντος καὶ µὴ ὄντος]. Similarly, 'recovered' ( he was healthy ) [ὑγιανεῖ] and 'will-recover' ( he will be healthy ) [ὑγίανεν] are not verbs but inflexions of verbs [ptῶσις ῥήµατος]. They differ from the verbs in that it (verbs) additionally signifies the present time [προσσηµαίνει χρόνον], they the time outside the present [τὰ δὲ τὸ πέριξ]. Ammonius indicates that the account of indefinite verbs is similar to that of indefinite names. 173 He claims that indefinite verbs are so named by Aristotle 172 Cooke (1938) translates this as Though they certainly have the time-reference and function at all times as predicates, I know of no recognized name , 25ff.

79 68 because [they] are not said of something definite. 174 This claim by Ammonius is somewhat confusing. It would seem that indefinite verbs are indefinite because they say something indefinite of a subject but not because they are said of an indefinite subject. Perhaps, Ammonius is focusing on the claim that indefinite verbs hold of both existent and non-existent subjects. 175 However, this would still seem odd. It does not follow that a non-existent subject is indefinite nor does it follow that since and indefinite verbs hold only of indefinite subjects. It appears that Ammonius has the predication wrong which is even more surprising given the connection he had just made between the account of verbs in Peri Hermeneias and the account of predication in the Categories. It seems clear from the passage that indefinite verbs predicate something indefinite. For example, is-not-well [οὐχ ὑγιαίνει] will not predicate any definite action or state. Not-well marks off a class of predicates that are heterogeneous. Before moving on to Ammonius s investigation of Aristotle s account of affirmations and sentences, there are two important passages on which Ammonius focuses that I also should consider: 16a10-19 and 16b Taken together, these passages can give us insight not only into names and verbs but also into Aristotle s purpose in discussing them. Ammonius indicates that, for Aristotle, all simple vocal sounds (i.e., names and verbs) resemble thoughts without composition or division , See Bäck (2000), 202 for a discussion of the complement class of indefinite verbs.

80 69 and are receptive of neither truth or falsity. 176 It is only when names and verbs are combined to form sentences that determinations of truth and falsity can be made. Ammonius, in the beginning of his analysis of chapter 4, points out the fact that Aristotle has been systematically working toward an account of sentences. Ammonius view is compatible with the programmatic statement at 16a1. In chapter 1-3, Aristotle has developed an account of the elements of sentences (i.e. names and verbs), and in chapters 4-6 he developed an account of sentences. This leads us to Ammonius analysis of Aristotle s account of sentences and affirmations. At 16b26ff Aristotle states: A sentence [λόγος] is a significant spoken sound (significant speech) [φωνὴ σηµαντικὴ] some part of which is significant (have meaning) [σηµαντικόν]- as an expression (as something) [φάσις], not as an affirmation [κατάφασις]. Ammonius claims that the reason Aristotle states that some parts of the sentence (i.e., names and verbs) are significant as expressions is to distinguish names and verbs uttered as parts of a sentence from those uttered independently. 177 Aristotle s later statements at 17a18-19 support this. 178 Ammonius also states that a further , , There Aristotle claims, (Let us call a name (nouns) or a verb simply (merely) (alone) [µόνον] an expression [φάσις], since by saying [εἰπεῖν] it one cannot reveal [δηλοῦντά] anything by one's utterance (one s speech) [τι τῇ φωνῇ] in such a way as

81 70 distinction is made by Aristotle between the parts of sentences that are expressions and the parts that are affirmations. Ammonius claims that compound sentences could have parts that are themselves affirmations. 179 Lastly, it is worth noting that Ammonius compares Aristotle s discussion here with Plato s Cratylus. 180 This gives us further evidence that some the early Greek commentators viewed Aristotle s discussion as being indebted to Plato s work in the Cratylus. Next, Ammonius considers Aristotle s statement at 17a1-4 that: Every sentence [λόγος] is significant (not as a tool {instrument of nature} [ὄργανον] but, as we said, by convention [κατὰ συνθήκην]), but not every sentence is a statement-making sentence (proposition) [ἀποφαντικὸς], but only those in which there is truth or falsity. Ammonius claims that Aristotle is refuting an argument that sentences are significant by nature. 181 According to Ammonius, the argument states that sentences are the tools or natural instruments of the vocal power within us. 182 The to be making a statement (make known/declare) [ἀποφαίνεσθαι], whether one is answering a question [ἐρωτῶντός] or speaking spontaneously [προαιρούµενον].) 178 This clearly supports Ammonius view , , Ammonius draws parallels between Plato s claim that names are the smallest part of a sentence and Aristotle s commitment to names being the smallest significant parts of sentences , 21-63, ,

82 71 argument claims that [e]very organ of a natural capacity is itself natural as well, just as the eyes, being organs of the optic capacity in us, are the work of nature not of art. 183 Therefore, the sentence is natural. He claims that Aristotle s solution is to claim that vocal power within us is natural (i.e. results from the natural operation of our organs) but the signification of names is conventional just as locomotion is natural, but dancing is [by convention]. 184 Ammonius goes on to indicate that, the dancer is able, by moving hands in this way or that, to signify, say, Achilles, representing by means of gestures now some, now others of his qualities. 185 Here, there is a tacit connection being made between Aristotle s discussion and Plato s discussion in Cratylus at 385b2-d1, where the truth/falsity of names is discussed. Ammonius notes that Aristotle begins to narrow the range of sentences he will consider to assertoric [statement-making] sentences and produces the following schema for them based on 17a8-10: The first [πρῶτος] statement-making sentence [λόγος ἀποφαντικὸς] is the (simple) affirmation [κατάφασις], next is the (simple) negation [ἀπόφασις]. 186 The others are single (one)[εἷς] in virtue of a connective (by conjunction) [συνδέσµῳ] , , , Cooke (1938) translates this, as A simple affirmation is the first kind, a simple negation the second of those propositions called simple.

83 72 This can be graphically depicted as follows: Figure 2 Statement-Making Sentences At 17a20-26, Aristotle claims that a simple statement affirms or denies something of something. An affirmation affirms something of something and a negation denies something of something. Ammonius argues that affirmation is prior to negation because adding the negative particle to what is affirmed forms a negation. 188 Therefore affirmation must be prior to negation. At 17a9, Aristotle claims that statement-making sentences must contain a verb (or an inflexion of a verb). Ammonius is unsure whether or not indefinite names can be used. Aristotle does not explicitly include indefinite verbs in his statement of the constituents of statement-making sentences. This leaves the possible role of indefinite names (in statement-making sentences) unclear. Ammonius, for his part, appears to rule out 187 Cooke (1938) translates this as The rest are but one by conjunction , 20 68, 9; 78, 26-79, 9; 80, 15-81, 1.

84 73 indefinite verbs in affirmations but leaves room for them in negations. 189 Lastly, Ammonius claims that Aristotle makes an important distinction between statementmaking sentences and names/verbs. Names and verbs are expressions (φάσες) that signify something but do not indicate truth or falsity. 190 Statement-making sentences, on the other hand, always indicate something true or false. They indicate one and only one thought , , 20ff. 191 Bäck (2000), 100.

85 Findings and Conclusions I am now in a position to examine the larger implications of the early Greek commentators work. One of the first things that I should recognize is the challenging nature of the text. As I shown, there were controversies over how to properly understand Aristotle s claims and their implications. One example of this is the dispute that arose regarding how to understand Aristotle s claims regarding πάθηµατα of the soul. The lack of accents in the received text as well as the ambiguity of some of Aristotle s claims forced the early Greek commentators to make interpretive decisions. Some of these decisions clearly distorted the meaning of text. As I indicate in the next chapter, similar challenges still confront modern commentators and those challenges require that interpretive decisions be made. In addition, I present some of the decisions made by modern commentators that affect the meaning of the text This leads to the next thing that should be noted. There were occasions when the early Greek commentators commitment to the harmonization of Platonic and Aristotelian philosophies caused them to make questionable interpretive decisions. The prime example of this is the early commentators commitment to the harmonization of Platonic and Aristotelian philosophy that caused them to incorrectly interpret the accounts in the Cratylus and Peri Hermeneias as being the same. Similarly, certain philosophical commitments may affect the ways in which modern scholars interpret the text of Peri Hermeneias. Being mindful of this is key to an accurate assessment of their views.

86 75 Lastly, the early Greek commentators recognized an over-arching architecture to Aristotle s work in Peri Hermeneias. They noticed that Aristotle systematically moves from an account of the elements of sentences (i.e., names and verbs) to an account of sentences and then to an account of a specific type of sentence (i.e., the statement-making sentence) that can be evaluated for truth/falsity. Moreover, their understanding of this architecture is key to their understanding of Aristotle s claims in Peri Hermeneias. As I evaluate modern interpretations of Peri Hermeneias, I should be mindful of this overarching program. Examining his claims outside of this framework runs the risk of creating a caricature of his views instead of an accurate characterization.

87 76 CHAPTER THREE: Modern Commentators and Peri Hermeneias 1-6 Introduction I am now in a position to examine contemporary scholars work on Peri Hermeneias. Much of the contemporary literature has centered on the first chapter of Peri Hermeneias. In particular recent scholarship has focused on Peri Hermeneias 16a3ff where Aristotle states: " Now spoken sounds [φωνῇ] (the things that are in the voice/sound) are symbols [σύµβολα] of affections in the soul [ἐν τῃ ψυχῇ παθηµάτων] and written marks [γραφόµενα] symbols of spoken (the things that are in the voice/sound) sounds. And just as written marks are not the same [τὰ αὐτά] for all men, neither are spoken sounds. But what these [ταῦτα] 193 are in the first place [πρώτως] signs [σηµεῖα] of - affections of the soul [παθήµατα τῆς ψυχῆς] - are the same [ταὐτὰ] for all; and what these affections are likenesses (representations or likenesses, images, copies) [ταῦτα ὁµοίωµατα] of - actual things [πράγµατα] - are also the same [ταὐτά]. These 193 Magee (1989), 8-12 points out that there was considerable controversy over whether ταῦτα and ταὐτά should be used in 16a6-8. This was discussed in section of this thesis.

88 77 matters have been discussed in the work on the soul and do not belong to the present subject." These roughly 9 lines in the Greek text are considered by some to be the heart of Aristotle s semantic theory and the key to understanding the claims made in the subsequent chapters in Peri Hermeneias. As I indicated in chapter two this is a philosophically and philologically rich and puzzling passage. Over time, commentators have offered a variety of differing interpretations of the aforementioned claims. Hans Arens describes this passage as having "incredibly far-reaching influence in the development of linguistic thought," 194 while J.L. Ackrill dismisses these claims as inadequate and elliptical. 195 Norman Kretzmann called this "the most influential text is the history of semantics." 196 David Sedley claims that "Peri Hermeneias is...one of the most misinterpreted texts in ancient semantics." 197 For these and other reasons, the account found in these lines has been the subject of a great deal of scholarly discussion and debate. 198 My goals in this chapter are to set out the details of the account contained in the lines above. This 194 Arens (1984), Ackrill (1963), However, Kretzmann (1974), 3 takes this to represent a theory of signification not a theory of meaning. This will be explained in more detail in chapter four. 197 Sedley (1996), Recent entries in the debate have been made by Modrak (2001), Wheeler (1995), Charles (2000), Butler (1997) among others.

89 78 sets the stage for the evaluation and reconstruction of this account in subsequent chapters. On first blush, one is confronted with the odd nature of some of the claims made in the account above. 199 There are two particularly troubling semiotic relations posited. First there is a relation posited between a significant sound/written mark and an internal state (παθήµα). No clear indication is given regarding the nature of this state. Is it a mental image, or a cognitive state, or something else? Also, the nature of the connection between the significant sound and internal state is not fully explained here. The second relation posited is between the aforementioned internal state (παθήµα) and an external object (πρᾶγµα). This relation is described as being one of likeness. But what does it mean for an internal state to be a likeness of an external object? Both of these relations are said to be 'the same for all', but how is one to understand this phrase? Having examined, in the preceding chapter, how the early Greek commentators wrestled with these questions, I am now in a position to examine recent discussions of these issues. 199 I make a number of assumptions that should be noted at this point. I follow Arens (1984), Wheeler (1995), Ackrill (1974), and Whitaker (1996) in assuming that Aristotle uses γραφόµενα and γράµµατα to refer to all written language not merely Greek (or any other particular language). I also assume, along with those just mentioned, that Aristotle uses φωναί to refer generally to the spoken sounds of human language. I also assume that there is no relevant difference between the phrase affections in the soul at 16a3 and the phrase affections of the soul at 16a6. Lastly, I join the consensus of scholars in assuming that Aristotle is referring to De Anima at 16a8.

90 79 In order to address the aforementioned questions, clarify the relations discussed above, and set the stage for the discussion of signification and semantic theories in chapter four, I explore the nature of the elements that constitute both relations. In section 3.1, I discuss (in turn) the sign and symbol relations. I critically examine the positions in the literature. Some commentators claim that sign and symbol pick out the same relation, while others claim that the sign relation is natural and symbol relation is conventional. In section 3.2, the focus shifts to the likeness relation and affections in the soul. Little is said in Peri Hermeneias regarding the nature of affections in the soul, but at 16a8 Aristotle states that "[t]hese matters we have discussed in the work on the soul and do not belong to the present subject." As I have indicated this is traditionally taken as a reference to De Anima, but it is unclear to which part of that work Aristotle is referring. I investigate the competing candidates for affections of the soul (παθήµατα). Next, in section 3.3, I discuss the actual things (πράγµατα) that are claimed to the objects of παθήµατα. Again, precious little is said explicitly about the πράγµατα in Peri Hermeneias. Fortunately, another work in the Organon discusses them at length. The classification schema of the Categories provides us with ontological schema to understand Aristotle's views on the world of existing things.

91 Sign and Symbol Both 'symbol' [σύµβολον] and 'sign' [σηµεῖον] are used by Aristotle, in this passage, to describe the relation between spoken sounds and affections in the soul. 200 Understanding Aristotle's use of these terms helps to clarify the nature of these relations. There are several positions in the literature regarding how to understand these terms. In Chapter two, I noted that the early Greek commentators considered 'sign' [σηµεῖον] and 'symbol' [σύµβολον] to be synonymous as conventional signs. 201 This position was handed down and dominated the literature on Peri Hermeneias until Kretzmann s seminal paper. Kretzmann (1974) identifies this position as part of what he calls the Received View (RV). The RV not only claims that sign and symbol are synonymous but also that the claims in the passage above establishes a set of transitive relationships with semantic import with sign and symbol playing a central role. 202 Kretzmann argues for a distinction between sign [σηµεῖον] and symbol [σύµβολον], which results in the denial of both the transitivity and semantic import of the claims above. Kretzmann claims that sign refers to a natural relation between affections in the soul and real things or affections in the soul and spoken sounds. Symbol refers to a non-natural or conventional relation between spoken sounds and written marks or the relation 200 Peri Hermeneias 16a2 and16a See section 2.2 of this thesis. 202 This is discussed fully in section of this thesis.

92 81 between spoken sounds and affections in the soul. 203 Signs are symptomatic indices of things. For example, smoke is a symptomatic index of fire. 204 Symbols, on the other hand, are conventional encoding mechanisms; they are rule governed embodiments of things that take place in another medium (i.e. the soul). 205 So, on Kretzmann s view written marks are encoding symbols of spoken sounds and spoken sounds are encoding symbols of mental impressions. 206 Most importantly, there are two ways of thinking about spoken sounds. In one sense, spoken sounds are symbols of affections in the soul and, in another sense, spoken sounds are signs of affections of the soul. 207 Kretzmann claims that the natural sign relation (of a spoken sound to an affection of the soul) is logically prior to the conventional symbol relation between a spoken sound and an affection of the soul. 208 To support this distinction between signs and symbols, Kretzmann cites Peri Hermeneias 16a Here Aristotle distinguishes between the sounds made by beasts and human language. Both the meow of cats and the ouch of humans are natural symptomatic signs, but only humans have language. 203 Kretzmann (1974), Kretzmann (1974), Kretzmann (1974), Ibid. 207 Here Kretzmann reads the ταῦτα as referring to spoken sounds only. 208 An analysis of this claim by Kretzmann and his rejection of what Wheeler (1995) calls the Semantic Primacy Thesis (i.e. the claim that spoken sounds signify primarily affections of the soul and secondarily real things) can be found in Section 4.2.

93 82 In response to Kretzmann many commentators have argued that sign and symbol are synonymous. Magee (1989), Modrak (2001) and Wheeler (1995) all point out that Kretzmann is correct in pointing out the conventional connotations of σύµβολον. However, these commentators (and others) appeal not only to the long tradition that views them as synonymous but also to the lack of philological support as reasons for rejecting Kretzmann s position. Both Wheeler (1995) and Modrak (2001) appeal to the lengthy tradition that views σηµεῖον and σύµβολον as conventional signs based upon social practice and agreement. 209 In addition, Magee (1989) indicates that there is a lack of philological support for Kretzmann s position. 210 He reveals that Kretzmann supplies insufficient evidence outside of Aristotle. 211 Only in one footnote does Kretzmann offer support for his position in the literature. 212 I should note that these considerations alone do not refute Kretzmann s position. I need further reasons to reject Kretzmann s claim that σηµεῖον should be understood as a symptomatic index of things. Wheeler (1995) indicates Aristotle s 209 Both Modrak (2001), 20 and Wheeler (1995), note the custom of calling the pieces of a coin that is broken, as a sign of a contract, σύµβολον. Wheeler (1995) also observes that Aristotle uses it refer to the halves of a tally stick. 210 Magee (1989), Ibid. 212 In his footnote 8, Kretzmann (1974) mentions Waitz, Brandt and Aubenque, but does not adequately explain how these references support his position.

94 83 explication of σηµεῖον in the Prior Analytics gives us reason to doubt Kretzmann position. According to Wheeler, Aristotle is committed to the following principle: Dsign: x is a sign of y if, and only if, when x is, either x is simultaneous with y or x precedes y or x is subsequent to y. 213 Wheeler correctly notes that, under this definition, all symbols are conventional signs and all signs are strongly connected with what they signify. 214 The implications of this interpretation are that key parts of Kretzmann s distinction between σύµβολον and σηµεῖον collapse. Applying Wheeler s reading to Peri Hermeneias 16a3ff, I see that every written mark is a sign of some spoken sound and that every spoken sound is a sign of some affection of the soul. 215 As I maintain in the next chapter, this reading importantly maintains the transitivity and reflexivity of the sign relation. In a similar vein, Muri argues (among other things) σύµβολον has connotations, which are derived from συµβάλλειν (to infer or conclude), that are strongly related if not synonymous with σηµεῖον. 216 Understood in this way, a σύµβολον is involved with inference and the understanding of that with which it is connected. 217 Also, the connection between the σύµβολον and its 213 Wheeler (1995), Wheeler (1995), Wheeler (1995), Magee (1989), Ibid.

95 84 significate can be conventional or natural. 218 Lastly, as Modrak (2001) points out Aristotle s preferred word for signification is σηµαίνω and he frequently uses σηµεῖον as a cognate. 219 Given this it is plausible to think of words, not as symptomatic indices, but as signifiers of mental content. On balance this evidence undercuts Kretzmann s position and gives us reason to accept the plausibility of the claim that σύµβολον and σηµεῖον are synonymous. 218 Magee (1989), Modrak (2001), 20.

96 Likenesses (ὁµοίωµατα) and Affections of the Soul (παθήµατα) 'Likenesses' (ὁµοίωµατα) is used by Aristotle to describe the relation between affections of the soul (ἐν τῃ ψυχῇ παθηµάτων) 220 and actual things (πράγµατα). 221 There has been a great deal of confusion concerning the proper way to understand this relation and its constituents. This confusion is only compounded by Aristotle s brief and unclear reference to De Anima at 16a Perhaps the best place to start to resolve this is to examine what Aristotle may means by likenesses (ὁµοίωµατα). As Wheeler (1995) reveals Aristotle discusses likeness in Categories III.8. Aristotle states that likeness is a function of qualities possessed by things. Any two or more things are like (ὅµοιος) and unlike (ἀνόµοιος) with respect to the category of quality. 223 As Wheeler indicates, this can be schematized as follows: If x is like y then there is some quality z such that x has 220 I follow Wheeler (1995) p195 in taking the phrases ἐν τῃ ψυχῇ παθηµάτων and παθήµατα τῆς ψυχῆς to be equivalent and interchangeable. 221 Peri Hermeneias16a I have shown in Chapter two that Andronicus thought that Peri Hermeneias was inauthentic due the vagueness of the reference to De Anima. Polansky and Kuczewski (1990) argue on page 54 that the reference here is not to some particular part of De Anima but to his basic account of cognition. This combined with Aristotle s supposition that the theory of cognition that undergirds his conception of ὁµοίωµατα was well known to his audience explains the vagueness of this reference. 223 Wheeler (1995), 205 and Categories 112a5-9.

97 86 z and y has z. 224 The implications of this are that παθήµατα resemble real things in that παθήµατα must some properties (qualities) in common with real things. Thus any candidate account of παθήµατα must explain how (or in what ways) παθήµατα resemble πράγµατα. This account of likeness is strongly tied to and informed by Aristotle s account of perception and thought in De Anima and Posterior Analytics. Examining these accounts not only shed light on Aristotle s conception of likeness but also help clarify the nature of the constituents of that relation (i.e. παθήµατα and πράγµατα). 225 However, there is general consensus in the literature that at minimum some of the keys to understanding παθήµατα are contained within De Anima. As Modrak (2001) points out Aristotle uses παθήµατα and πάθος for a number of emotional and cognitive states that include thinking, perceiving, loving, etc. 226 The natural question now arises with which πάθος should the internal state mentioned in Peri Hermeneias 16aff be identified. It is clear from the discussion above that any putative candidate account for the πάθος I seek should cohere with Aristotle s account of perception and thought. 224 Wheeler (1995), 205. Similar claims can be found in Polansky and Kuczewski (1990), Modrak (2001), and Charles (2000). 225 The likeness relation and the nature of παθήµατα will be discussed in this section of this thesis. The πράγµατα (and its role in the likeness relation) will be fully explored in section 3.3 of this thesis. 226 Modrak (2001), 221 cites De Anima 403a3 25.

98 87 The available candidates for affections of the soul appear to be mental images and thoughts. There are two candidates for the παθήµατα I seek: φαντασία (imagination) and νόηµα (abstract thought). Since a connection is made, Peri Hermeneias 1, between spoken sounds and παθήµατα, I begin by following Wheeler (1999) and Modrak (2001) in examining Aristotle s distinction between sound (ψόφος), voice (φωναί), and speech (λόγος ὁ διἀ τῆς φωνῆς). In De Anima II.8, Aristotle discusses sound and hearing. He indicates that all sound is produced when something strikes against a mass of air and set that mass in continuous motion to some organ of hearing. 227 At 420b29ff, Aristotle claims that voice is a species of sound produced by an animal. However, he qualifies this claim by stating that not every sound made by an animal ought to be considered voice. 228 As Modrak (2001) points out a similar division is found in Historia Animalium. 229 In Historia Animalium IV.9, 230 Aristotle employs a trichotomous distinction 231 between sounds an animal makes by emitting internal air (e.g., humming, buzzing, squeaking, or moaning), voice (φωνή) which requires lungs and pharynx, and articulated voice/speech or language (διάλεκτος). 227 De Anima 420a De Anima 429b Modrak (2001), 220 points out that voice is mentioned more frequently in Historia Animalium than in De Anima b29 536a Modrak (2001), 220 notes that all the elements of this trichotomous distinction are defined physiologically.

99 88 Figure 3 A wide class of creatures is capable of producing sound either by emitting internal air or by the air friction produced by their body parts. 232 Aristotle recognizes that some animals produce voluntary sounds that have natural significance for other members of their species. 233 These are instances of voice and require lungs and a windpipe. 234 Just as not all animals are capable of voice an even smaller number of animals are capable of articulated speech. One necessary condition for articulate speech/voice is that the tongue be able to move freely. 235 Another necessary condition is physical maturation. Aristotle claims that full 232 Historia Animalium 535a Historia Animalium 536a Historia Animalium 536a This is why dolphins are not capable of articulated speech while many birds are able to produce articulated speech.

100 89 control of the tongue is only acquired through physical maturation. 236 There are other factors that shape an animals production of articulated voice, such as location and nurture. 237 Nurture or early experience conditions articulated voice. Aristotle states that birds raised by foster parents will produce different song than that of their parents. 238 The last necessary condition is that voice must have soul in it and must be accompanied by an act of imagination (φαντασία). 239 According to Aristotle, language is the articulation of voice by the tongue, lips, and larynx. 240 Despite noting the ability of other animals (e.g., birds and) to produce articulated voiced, Aristotle claims that language is unique to human beings. 241 This presents us with a puzzle. On first blush, humans appear to lack a distinguishing feature that would justify this difference. Humanity s unique ability to produce language cannot be explained/justified on a physiological basis because there are instances where Aristotle treats the articulated voice of other species as 236 Historia Animalium 536a Modrak (2001), Historia Animalium 536b De Anima 420b Historia Animalium 534b31 535a Historia Animalium 536b2. This is also reiterated in De Generatione Animalium 768b20-21, where Aristotle states, nature has given to him in the highest degree because he alone of animals makes us of speech and the voice is the material of speech.

101 90 being equivalent to human language. 242 What is needed here is an account that provides a non-arbitrary basis for Aristotle s restriction of language to humans. The account needed here explains the reference to Aristotle s psychological works at 16a8 and clarify the παθήµατα I seek. I am now in a position to situate the candidates for the παθήµατα within Aristotle s overall account of the soul and evaluate them. Aristotle s account of cognition is embedded in his general account of the soul. 243 In De Anima, Aristotle s discussion focuses on the capacities of the soul moving from nutrition to perception to mind. 244 In addition to the capacities listed, Aristotle also discusses two other capacities: desire and imagination (φαντασία) Modrak (2001), 221 cites 610a17 where Aristotle points out that a male elephant will flee when he hears the voice of another male who has defeated him. This indicates that, for animals of the same species voice is a natural sign of the state, location, and intent of another animals. Modrak (2001), 220 also points out that Aristotle makes a similar claim at De Anima 435b24 where he claims that animals posses hearing so that something may be signified to one by another animal. 243 Aristotle s investigation of the soul is found mainly in De Anima but related discussions are found in De Sensu and De Memoria. 244 We should note that Aristotle recognizes a hierarchy of living things that follows the psychic capacities listed. All living things have nutrition (De Anima II 4, 415a24-25), but only animals have perception and only humans have mind. I will appeal to this hierarchy later in this section to explain Aristotle s restriction of language to humans. 245 Imagination is discussed in De Anima III.3 and desire is investigated in De Anima III.9-10.

102 91 Let us first turn to νοήµατα (thoughts). The noetic capacity is discussed in De Anima III Aristotle the noetic capacity as the part of the soul by which it knows and understands. 246 This definition betrays Aristotle s supposition that it is in our very nature to desire knowledge and understanding. 247 Aristotle claims that the possession of the noetic capacity is essential to being human. To explain the functioning of the noetic capacity Aristotle draws analogies to the perceptive capacity. Just as perception involves the reception of a sensible form by a suitably qualified sensory organ, thinking involves the reception of an intelligible form by a suitably qualified intellectual faculty. 248 The mind is not actually anything until it begins to conceive of an object of thought. Thinking occurs whenever some suitably disposed mind is made like its object by being affected by that object. 249 According to this model, thinking consists of a mind's becoming enformed by some object of thought in such a way that a likeness of the object of thought becomes present in the thinker s mind. This results in the same form being common to both the object of thought and to the thinker De Anima 429a9 10; cf. 428a5; 432b26; 434b Metaphysics 980a21; De Anima 414b18 and 429a De Anima 429a13 18; Whitaker (1996), and Modrak (2001), present a similar understanding of the functioning of the noetic capacity. 249 Ibid. 250 At 430a3-5 Aristotle claims that which thinks and that which the thought is about are the same. Also, at 431a1ff, he indicates that knowledge is the same as the thing known.

103 92 The fact that thoughts are likenesses of objects of thought, in that they share the same form, has led Whitaker (1996) among others to identify νοήµατα with the παθήµατα in Peri Hermeneias that are likenesses of actual things. 251 In addition Whitaker (1996) and Magee (1989) both point to the analogy between language and thought at Peri Hermeneias 16a Here Aristotle claims simple names and verbs are like simple thoughts in that they cannot be evaluated for truth or falsity until they are combined. Both Magee (1989) and Whitaker (1996) argue that Aristotle s appeal to νοήµατα here is deliberate. 253 They claim that this analogy shows that Aristotle is clearly interested in the relation between words and thoughts, and not that between words and mental images [φαντάσµατα]. 254 In addition, Whitaker (1996) points to Aristotle s discussion of verbs in the third chapter of Peri Hermeneias as evidence. At 16b20 Aristotle points out that both the speaker and the hearer grasp in thought the utterance of a verb by itself. Internal to this passage is there an indication that φαντάσµατα are at work. Instead, Aristotle once again mentions thoughts. With respect to the schema in Peri Hermeneias 1, written marks would be signs of words and words would be signs of νοήµατα. Νοήµατα would be the likenesses of objects of thought (i.e. the formal characteristics of πράγµατα) and the 251 Magee (1989), argues in a similar vein that the παθήµατα should be understood to be as first and simple thoughts. 252 Whitaker (1996), and Magee (1989), Ibid. 254 Whitaker (1996), 15.

104 93 meaning of words/written marks. For example, thinking about a cat requires conceiving of a likeness in one s mind that is determined by the formal characteristics of cats (i.e. what a cat is). Just like cats themselves, the thoughts of cats would be the same for all. All cats are cats since every cat shares the same form with all other cats. Because thoughts of cats are formal likenesses of cats themselves, both the thoughts of cats and cats themselves will have the same form. Since on this picture words are conventional signs of νοήµατα, the word cat would have as its content the thought of cat. As such the meaning of the word cat would be the formal characteristics that are content of thoughts about cats and definitional of the type cat. While I could adopt a convention to use the word goat instead of cat for thoughts about cats, I cannot by fiat or convention chose which thoughts to use with which things in the world. In addition, this view coheres with the account of articulate speech and language presented earlier in this section. On this view, the restriction of language to humans is justified because humans are the only ones capable of abstract thought. So it seems that I have some good evidence for νοήµατα being the παθήµατα I seek. However, there are some serious problems when I attempt to fully place νοήµατα in the role of the παθήµατα in Peri Hermeneias 1. While νοήµατα are likeness of objects of thought, this does not quite seem to fit the role of the παθήµατα in Peri Hermeneias 1. The actual things of which παθήµατα are likenesses are concrete particulars grasped via perception, but νοήµατα are universals abstracted from particulars. It is the perceptual faculty that apprehends

105 94 the sensory content of concrete particulars that exist independently of cognitive activity and act upon the sensory organs. As such, the perceptual faculty is connected directly to the world. The objects of the noetic capacity, on the other hand, are not directly grasped through sensory content. There seems to be a clear disconnect between the content of thoughts and the content of perceptions. Another problem centers on the account of speech presented earlier in this section. 255 If νοήµατα are the παθήµατα I seek, then it is hard to understand Aristotle s requirement that every act of speech is accompanied by φαντασία. 256 As Modrak (2001) points out, it would seem that φαντασία has no role to play. 257 If thoughts are the meanings of words, then it seems that at best φαντάσµατα are unnecessary add-ons. Words and written marks have νοήµατα as their meanings. If Aristotle s claim is to be taken seriously, then some adequate account of this requirement is needed. Let us now turn to the other candidate for the παθήµατα I seek: φαντασία. Aristotle apparently views imagination, which includes capacities such as imagining, remembering, and dreaming, as a kind of subordinate faculty, connected in various ways with the faculties of nutrition, perception, and thought. However, the connection between φαντασία and perception is the strongest. 255 See page of this thesis. 256 De Anima 420b Op. Cit., 245.

106 95 In order to better understand this connection I now turn to Aristotle s account of perception. In De Anima, Aristotle spends a good amount of time discussing perception. He claims among other things that perception is what distinguishes animals from plants. 258 He also claims that the process of perception involves a type of alteration 259 where like is affected by like. 260 As Aristotle states at De Anima 418a3-6: That which perceives is potentially what the perceptible object actually is already, as I have said. When it is being affected, then, it is unlike the object; but when it has been affected it has been made like the object and has acquired its quality. In other words, perception involves things that are capable of acting and things that are capable of being acted upon. Both of these must be capable of affecting/being affected in the appropriate way. This involves the affected thing becoming like the thing that affects it. In order for this to happen, the affected thing must have the requisite active capacities. The perceptible object affects the percipient by en- 258 This claims is made in De Sensu 1, 436b9-12 and De Anima. 259 A full discussion of the claim that perception involves alteration is beyond the scope of this dissertation. I will only note the distinction that Aristotle makes at 417b3 418a6 between two types of alteration. One type of alteration involves the destruction of one contrary by another. While the other type concerns preservation of a thing as I moves from 1 st actuality (2 nd potentiality) to 2 nd actuality. Perceptual alteration is like the second of these two types of alteration. 260 De Anima 416b32-35.

107 96 forming its sense organs (i.e. the sensible form of the perceptible object becomes en-mattered in the thing that is affected). Let us examine this account in more detail. Higher animals perceive through the five basic senses and the common sense. Each basic sense involves the capacity to grasp a certain kind of object. The particular object, which is focus of each sense, is defined by a pair of opposites (e.g., light-dark, hot-cold, high-low, etc.). 261 Perception is the reception of the sensible form (the form without matter). This sensible form is the logos of opposite qualities. Perhaps an example here will shed some light on Aristotle s position. Let us consider a person in a sauna who regrettably decides to pick up one of the heating stones. The hot stone acts on the individual s hand by heating it. The form of heat passes from the stone to the individual s hand. The form of heat is taken on by the matter (i.e. flesh) of the person s hand. While the form is not without matter in this case, in cognition the form is experienced without incorporating it into matter. The individual s hand is hot but their sense of touch only perceives the heat. While the sense organ materially receives the form, 262 the active sense has the form immaterially. The account of perception that Aristotle provides has a causal basis. The world experienced by the perceiver is causally connected to the world as it is. A material causal chain connects the perceiver to the world via the sensible form. The heat the 261 Modrak (1987), That is to say that the sense organ incorporates the form into its matter. The sense on the other hand takes on the form without incorporating it into matter.

108 97 perceiver experiences is causally connected to the world itself. The heat occurs in the world and in the organ itself. Aristotle states that there are objects that have a role in perception other than the proper object of each sense. There are objects that are common sensibles (i.e. movement, number, figure, etc.) and those that are incidental sensibles (e.g., people, buildings, etc). 263 Common sensibles are apprehended καθ αυτά by the common sense when perception takes place through more than one sense modality simultaneously. 264 The perception of them is dependent on the percipients perception of the proper objects of those sense modalities. 265 For example, the common sensible of movement when grasped by sight and hearing requires perception of sound and color. Incidental sensibles, on the other hand, do not directly have an effect on any sense organ as common or proper sensibles do. 266 Incidental sensibles are not perceived in themselves (καθ αυτά), but are incidental (κατὰ συµβεβηκός) to the perception of some other sensible. 267 For example, the 263 De Anima 418a Modrak (1987), Modrak (2001), 225 and Modrak (1987), 64. Modrak indicates that common sensibles are perceived κατὰ συµβεβηκός when they are perceived only by an individual sense. They are perceived καθ αυτά when they are perceived by the common sense. The common sense emerges as a distinct perceptual capacity when the multiple senses combine to form a single perceptual act. 266 Modrak (2001), As Modrak (2001), 225 points out this is not to say incidental sensibles are inferred. While experience and memory have a role, the perception of incidental

109 98 individual who has the requisite experiences will perceive the white thing as a particular person Callias. 268 The perception of an incidental sensible involves sensible features that are perceptible καθ αυτά (i.e. proper and common sensibles). 269 Proper and common sensibles are not only vehicles for the perception of incidental sensibles but also they are vehicles for all other perceptions/functions of the perceptual faculty including apperceptual awareness and dreams. 270 Like the perception of incidental, proper, and common sensibles, apperceptual awareness is involves sensible features perceptible καθ αὐτα. Aristotle states in De Anima III.2 that each sense is aware of its operation (e.g., the awareness of touching belong to touch). 271 In addition, apperceptual awareness involves the coordinated functioning of the senses (in a way similar to the common sense). Aristotle claims that I am able to discriminate between the proper objects of each sense through a capacity that is common to all. 272 For example, I can tell the sweet from the white not by appealing to some sense other than that which I use to sensibles is spontaneous. Modrak (2001) p 225 juxtaposes the case of the perception of an incidental object with the case of someone reasons based on beliefs that the approaching object is a person known (i.e. Callias). The former involves inferring that the object approaching is Callias while the latter involves the spontaneous object perceived as a particular person. 268 De Anima 418a Modrak (2001), 225 and Modrak (1987), Ibid. 271 De Anima 425b12-23 and De Somno 455a De Anima 426b

110 99 sense either the sweet (i.e. taste) or color (i.e. sight). None of the special senses has the capability to make this judgment because each sense perceives its special object καθ αὐτα and can only perceive other sensibles κατὰ συµβεβηκός. Therefore, these judgments must be made by some other capacity common to all special senses. Before I move on to Aristotle s account of φαντασία, it is be helpful to highlight the parts of Aristotle s account of perception that I bring forward. First, it should be noted that the world as perceived by the percipient is causally connected to the world as it is in itself. Second, sensory content of is the essential content of the perceptual faculty. As I indicated sensory content is vehicle that facilitates the various functions of the perceptual faculty. Lastly, I showed that the perceptual faculty is capable of making judgments. This is shown in the perceptual faculty s ability to make cognitive discriminations between proper sensibles. Now that I have examined key features of Aristotle s account of perception, I am in a position to examine his account of φαντασία. Aristotle s initial definition of φαντασία is that in virtue of which I say a φάντασµα arises in us. 273 A φάντασµα is an image, appearance, and/or phantom that results from sense perception. These images remain in sense organs and have sensuous features (e.g., sound, color, shape, etc.). 274 Φαντασία is given a broad range of application to the activities involved in thoughts, dreams, and memories. In De Anima III.3, Aristotle claims that φαντασία is a movement resulting from an actual exercise of a power of 273 De Anima 428a De Anima 431a15.

111 100 a sense. 275 As such φαντασία is strongly connected to perception. Aristotle explicitly states that the faculty of φαντασία is the same faculty as the faculty of perception but they differ in essence. 276 The objects of φαντασία are the same as the objects of perception: 277 Figure 4 Objects of Perception and Φαντασία Objects of Perception Objects of Imagination (Phantasia) Common Proper Incidental Common Proper Incidental Aristotle points out that the φαντάσµατα produced by the different modes of sensory activity will differ. 278 Since φαντάσµατα are necessarily similar in character to the sensation that produces them, 279 those derived from the perception of special objects are never in error. 280 Those that are come from the perception of incidental objects may be in error. 281 For example, while the perception that there is something white before us cannot be false, the perception that what is white is Callias may be 275 De Anima 429a2-3 and De Insomniis 459a De Insomniis 459a16-18 and De Anima 432a31-b De Anima 428b De Anima 428b De Anima 428b De Anima 428b De Anima 428b19-21.

112 101 false. 282 Lastly, the φαντάσµατα that arise from the perception of common attributes (e.g., movement, etc.) admit of the greatest amount of sense-illusion possible. 283 However in De Anima III.3, Aristotle takes pains to distinguish φαντασία from both perception (αἴσθησις) and mind. 284 Aristotle distinguishes φαντασία from perception (αἴσθησις) on three main grounds. 285 First, he indicates that φαντασία produces images when there is no perception, as in dreams. 286 Second, he points to the biological fact that φαντασία is lacking in some lower animals, even though they have perception. Third, he claims that perception is always true whereas imagination can be false. 287 Aristotle also denies that imagination can be identified with mind or belief, or any combination of belief and perception. 288 Overall φαντασία can be conceived as a faculty in humans and most other animals, which produces, stores, and recalls the images used in a variety of cognitive activities, including those, which motivate and guide action. 289 More relevant to or present investigation, φαντασία can be understood as the awareness of sensory content under conditions that are not conducive to veridical 282 De Anima 428b De Anima 428b De Anima 414b33-415a De Anima 428a Ibid. 287 De Anima 428a De Anima 428a16-b De Anima 429a4-7 and De Memoria 450a22-25.

113 102 perception. 290 In developing the distinction between αἴσθησις and φαντασία Aristotle appeals to the difference between standard and nonstandard conditions. 291 The overriding concern here is whether or not the object is being perceived accurately. When the object is being perceived correctly that is an instance of αἴσθησις. If the perceptual conditions are not conducive to accurate perception, such when the object is at the edge of the perceptual organ s range or the organ is not healthy, then there is an instance of φαντασία. I can now turn to the ways in which φαντασία satisfies the likeness criterion and fits other characteristics of the παθήµα I seek. Φαντασία clearly satisfies the criterion that the παθήµατα are likeness of actual things. As I pointed out earlier, sense perception involves the sensory organ taking on the perceptible form of an external object. Φαντάσµατα are images with sensuous features that result from sense perception. The sensuous content of a φαντάσµατα is a complex of sensory characters [properties] that the external object possesses in relation to the percipient. 292 Given our earlier definition of likeness, 293 it is evident that 290 Modrak (2001), and Modrak (1987), Modrak (2001), 229 Standard conditions refers to a total perceptual situation or environment in which the perceptual object is apprehended. This includes but is not limited to the proper health and function of the sense organ, the perceptual object being within range of the organ, etc. 292 Modrak (2001), See page of this thesis.

114 103 φαντάσµατα are likeness of actual things (πράγµατα), in that φαντάσµατα have sensible qualities (properties) in common with actual things (πράγµατα). Further support for φαντάσµατα as likeness can be found in De Memoria. 294 Aristotle defines memory as the having of an image [φαντάσµατος], related as a likeness to that of which it is an image. 295 The φαντάσµατος is able to refer to the past sensory event (or object) because it is like the event. It is the similarities between the sensuous content of the φάντασµα and the past sensory event (object) that are of importance here. As Modrak (2001) points out, it is this same likeness that explains how a mental state can refer to an extra-mental referent. 296 The role of φαντάσµατα in memory also has important connections to Aristotle s epistemology. In Posterior Analytics II.19, a connection between παθήµατα, perception, and mind independent reality is developed. 297 Aristotle indicates that perception gives rise to a percept (αἴσθηµα) in the soul. 298 Αἰσθήµατα that are retained by memory (µνήµη) become memories (µνήµαι). 299 Multiple memories of the same thing give rise to a single experience (ἐµπειρία µία τῳ ἀριθµῳ). 300 Lastly, 294 Modrak (2001), De Memoria 451a Modrak (2001), Both Wheeler (1995), 204ff and Modrak (2001), 21 cite this passage as a part of their arguments regarding the connection between παθήµατα and perception/cognition. 298 Posterior Analytics 99b Posterior Analytics 99b39 100a Posterior Analytics 100a23 9.

115 104 many individual experiences generate a principle (ἀρχή). As Wheeler (1995) points out that all of these affections in the soul are derived from perception. 301 The role of φαντάσµατα in memory and the acquisition of first principles, explains important aspects of Aristotle s account of speech and voice. As I noted earlier Aristotle, through a series of distinctions, ultimately restricts language to humans. This narrowing of language runs parallel to the narrowing of active intellect to humans. Aristotle indicates in De Anima that voice has soul 302 in it and that voice must be accompanied by an act of φαντασία. 303 There is good evidence for thinking that φαντάσµατα seem well suited to be the παθήµατα in Peri Hermeneias 1. The παθήµα is said to be a likeness of πράγµατα and φαντάσµατα fulfill this role well. However, there are several problems that arise when I place φαντάσµατα further into the role of the παθήµατα in Peri Hermeneias 1. The παθήµατα is claimed by Aristotle to be symbolized by words/written marks. This has traditionally been taken to indicate that the παθήµατα are the meanings of words/written marks. If I place φαντάσµατα in that role then initially the account looks promising. As Modrak (2001) points out the meaning of the proper name Dante would be the sensory content contained in some φάντασµα (i.e. the image of a brown Dachshund). 304 Likewise, the meaning 301 Wheeler (1995), 204 referring to Posterior Analytics 87b28-88a De Anima 420b De Anima 420b Modrak (2001), 236.

116 105 of the word dog would be established by φαντάσµατα of Dante and other dogs. 305 The problem with this account is that it may explain conventional definitions of dog, but it cannot provide a satisfactory scientific definition. The scientific or real definition of dog cannot be composed only of sensible characteristics derived from perspectival perceptions of concrete individuals. 306 Many mass terms and common nouns have characteristics in their real definitions that may not be found in the sensory content of an individual s φαντάσµατα. The heart of the problem lies in the fact that φαντάσµατα appear to be limited to only representing particulars. Two additional problems crop up for this account. The first centers on correspondence between παθήµατα and words/written marks in Peri Hermeneias 1. The general schema presented in this chapter seems to imply a one to one correspondence between παθήµατα and words/written marks. Φαντάσµατα seem to be a poor match here. One of the positive aspects of φαντάσµατα is their ability to represent a range of sensory content up to and including complex states of affairs. This will not conform to a one to one correspondence model. As several recent commentators have pointed out a semantic theory that maps discrete semantic units onto parts of a mental representation is fraught with problems. 307 Not the least of which is how such an account would work for verbs, adjectives, and other linguistic 305 Ibid. 306 Modrak (2001), 236 and Modrak (2001), ; Ackrill (1963), ; Kretzmann (1974), 8-9.

117 106 items. However, as Modrak (2001) and Wheeler (1999) point out, these difficulties alone are not sufficient to show that Aristotle does not have a theory of this type. 308 The other problem focuses on the depiction of language in chapters 2-4 of Peri Hermeneias. Aristotle claims that significant sounds by themselves are neither true nor false. 309 It only when a ὄνοµα is connected to a ῥῆµα that something true or false is asserted. The challenges that this presents for an account of φαντάσµατα as the παθήµατα I seek is how to give a satisfactory explanation the ways in which mental representation such as φαντάσµατα can not only map onto the complex structure of language but also represent the combinatorial nature of those complexes. It one thing to provide an adequate account of the individual elements of a sentence like Agaphto sits next to Dante on the loveseat. Each word would have some correlated φάντασµα which would be its meaning. However, it is another thing to explain how the content of the sentence, which arise from the combination of these elements, is itself represented. Is there some φάντασµα that contains the content of a sentence? If so, then Aristotle seems to be committed to a picture theory of meaning that has been widely criticized. 310 Do a series of φαντάσµατα combine to form some more complex φάντασµα? If so, then a 308 Modrak (2001), 236 and Wheeler (1999), 207. This question is fully addressed in Section of this thesis. 309 Peri Hermeneias 16a For discussion of the problems with picture theories of meaning with respect to Aristotle s claims in Peri Hermeneias see Arens (1984), 31; Ackrill (1963), 113; Whitaker (1996),

118 107 coherent account must be given regarding the mechanics by which the content of the individual φαντάσµατα combine to form the singular content of a sentence. At first blush, it seems that φαντάσµατα lack this ability. 311 As I have shown both candidates have satisfied some of the requirements for the παθήµατα in Peri Hermeneias. However, serious problems arise when I attempt to fully place either candidate in the role of the παθήµατα. Since neither of the available candidates can provide a fully satisfactory account, I may be tempted to end my search without resolution. However, another option remains available. I can attempt to combine the accounts for the candidates to provide a satisfactory account of the παθήµατα I seek. There appear to be good prima facie reasons to believe that such an account will work. First, as I have found, the each of the candidates satisfies some part of the role attributed to the παθήµατα in Peri Hermeneias. Φαντασία clearly satisfies the likeness aspect and νοήµατα fits the structure and content aspects. In addition, as Modrak (2001) points out, all cognitive acts are an exercise of one or both of these faculties (i.e. the noetic and the perceptive). 312 Lastly, both of these faculties function in similar ways in that they are both potentially what their respective objects are actually. There is a major obstacle for an account that combines φαντάσµατα and νοήµατα to overcome. The objects of φαντασία are sensible particulars and the 311 Modrak (2001), 242 and Whitaker (1996), Op. Cit., 234.

119 108 content of φαντάσµατα are sensory characteristics while the objects of the noetic capacity are intelligible forms and the content of νοήµατα are universals. If this account is going to succeed then the rift between the content of νοήµατα and φαντάσµατα must be bridged. A promising route is can be found in Posterior Analytics and De Anima. 313 At 100a15-17, Aristotle states that while one perceives the sensible particular, the perception is of the universal. This indicates that there may be a way in which the content of φαντάσµατα connect to universals. Additional evidence in support of a combinatorial view can be found in De Anima III.8 where Aristotle states that nothing exists separate from sensible magnitudes. 314 Aristotle indicates here that this holds true for the objects of thought. 315 Lastly, at 432a12-14, Aristotle states that while νοήµατα are not to be conflated with φαντάσµατα, νοήµατα necessarily involve φαντάσµατα. It is this connection between φαντάσµατα and νοήµατα that is at the heart of my account of the παθήµατα in Peri Hermeneias 1. As I stated earlier φαντάσµατα are likenesses of existing things in that their sensuous contents in common with πράγµατα. 316 Also since, as I have pointed out, the various objects of φαντασία give it a wide capacity for representational flexibility. 317 This resolves the concern that φαντάσµατα are limited to only 313 Modrak (2001), a a See page of this thesis. 317 See page 101 of this thesis.

120 109 representing sensible particulars. As Modrak (2001) indicates, each sense perceptible particular is simultaneously a token of a number of types. 318 For example, Dante is simultaneously a dachshund, a dog, an animal, a living thing, and a short-legged quadruped. He is also shorthaired, ruddy, and more. Dante holds some of these types essentially while he holds others accidentally. All of these types (both the accidental and the essential) are exemplified in the φαντάσµατα that is a likeness of an extra-mental object. Extra-mental objects are complexes of form and matter. Similarly, essences inhere in φαντάσµατα in such way that φαντάσµατα serve as an internal model for the external object. 319 In this way νοήµατα can be seen as particular way of grasping the content of φαντάσµατα. Thought involves ignoring individuating characteristics and properties in a φάντασµα. 320 This is accomplished by the active intellect. In De Anima III.5, the active intellect is described as the efficient cause of thoughts in the passive intellect. 321 The active intellect is also described as maker of all things. 322 Lastly, an analogy is made between the active intellect and light. 323 Aristotle states that light is the activity of the transparent as such. 324 It is a static condition. 325 It is what allows colors to 318 Op. Cit., Modrak (2001), Metaphysics 1078a De Anima 430a De Anima 430a12 and 430a De Anima 429a3-5 and 430a De Anima 419b9-11.

121 110 produce sensation and so make themselves actual colors, by being seen as they actually are. 326 It makes the potential colors actual by being active. Without light color remains invisible. 327 When light is active in this way, its activity pervades the whole medium between object and percipient. This activity, in the medium, allows for color to impact the perceptual organ of the percipient and it results in the percipient perceiving color. 328 When this analogy is applied to the active intellect, I see the active intellect is the cause of thoughts in the passive intellect. The relevant universal will exist actually (not merely potentially) when there is appropriate activity in the active intellect. 329 The analogy runs as follows: Figure 5 The process for perception of color 330 ability of the object to change the transparent change in the transparent when active perception of color 325 De Anima 418b De Anima 419a De Anima 419b De Anima 419b Charles (2000), Charles (2000), 131.

122 111 Figure 6 The process for thoughts of universals 331 ability of the object to 'modify' the active intellect activity of the universal in the active intellect thought of object or universal (in the passive intellect) Both colors and universals exist independently of us. Just as light makes potential colors actual through its activity, the active intellect makes universals actual through its activity. The universals contained in the content of φαντάσµατα are made actual (in the passive intellect) by activity of the active intellect. In this way, νοήµατα can be seen as the actualization of the intelligible content potentially contained in φαντάσµατα. Support for this account can be found in De Anima. At 431b2, Aristotle states that the thinking faculty thinks the forms in φαντάσµατα. Additionally support for this account can be found in De Memoria 450a1-10, where Aristotle explicitly states that thinking is impossible without φαντάσµατα. Aristotle s explanation of this claim draws an analogy to the use of φαντάσµατα in abstraction performed as a part of geometrical demonstrations. In this process the intelligible 331 Charles (2000), 133.

123 112 form is separated from its material substrate. Aristotle indicates that a φάντασµα is used by the geometer to think about the abstract object (i.e. the triangle) as quantitative. 332 Certain accidental aspects of the φάντασµα are ignored. So that even though the geometer envisages the object as quantitative, she thinks of it in abstraction. 333 I have seen that the role of φαντάσµατα in thinking is evidenced in several places in Aristotelian corpus. Φαντασία has a role in language, perception, and cognition. Φαντάσµατα are the vehicles that connect the intelligible and sensory content of real particulars with our perceptual and noetic faculties. I have also seen that φαντάσµατα are flexible enough to particular object or a past event or abstract/universal concept. In the strictest terms, the παθήµατα of Peri Hermeneias 1 are φαντάσµατα. 332 De Memoria 450a De Memoria 450a6-7.

124 Real things (πράγµατα) Introduction In this section, I examine the actual things (πράγµατα) that are claimed to the objects of παθήµατα. Precious little is said explicitly about the πράγµατα in Peri Hermeneias. Fortunately, another work in the Organon discusses them at length. The classification schema of the Categories provides one with ontological schema to understand Aristotle's views on the world of existing things. Among the things Aristotle discusses in the Categories are the division of being (τἅ ὄντα) into four types, the ten-fold division of things said (τἅ λεγόµενα), and semantic relations. Early in the Categories, Aristotle presents a list of ten things said: Of things said without combination, each signifies either (1) a substance or (2) quantity or (3) qualification or (4) a relative or (5) where or (6) when or (7) being-in-a-position or (8) having or (9) doing or (10) being-affected. 334 This is a list of the ten fundamentally different kinds of things that there are to be said. The first category (i.e. substance) is the most important in Aristotle s ontology and, likewise, it is important for the account I develop. Substances are, for Aristotle, 334 Categories 1b25-28.

125 114 the fundamental entities. To see why this is the case, one has to understand what Aristotle claims about predication in the Categories. In the second chapter of the Categories, Aristotle presents a four-fold distinction of being (τἅ ὄντα): Of things there are: (a) some are said of a subject but are not in any subject. For example, man is said of a subject, the individual man, but is not in any subject. (b) Some are in a subject but are not said of any subject. (By in a subject I mean what is in something, not as a part, and cannot exist separately from what it is in.) For example, the individual knowledge-of-grammar is in a subject, the soul, but it is not said of any subject; and the individual white is in a subject, the body (for all color is in a body), but is not said of any subject. (c) Some are both said of a subject and in a subject. For example, knowledge is in a subject, the soul, and is also of a subject, knowledge of grammar. (d) Some are neither in a subject nor said of a subject, for example the individual man or the individual horse for nothing of this sort is either in a subject or said of a subject. 335 Aristotle distinguishes two fundamental relations: being said of a subject and being present in a subject. They cut across all ten categories. What is said of a subject are universals (i.e. secondary attributes) while things not said of are particulars. For 335 Categories 1a20-1b7.

126 115 example, man is said of Socrates and animal is said of horse. The said of relation involves ontological classification and its relata belong to the same category. The relation is also transitive such that if A is said of B and B is said of C then A is said of C. For example, animal is said of man and man is said of Rodmon; therefore, animal is said of Rodmon. What is present in a subject is accidental to that subject while what is not present in is non-accidental. What is present in a subject is accidental in that it does not belong as a part of the subject and cannot exist independently of that subject. 336 Here relata belong to different categories. For example, this grammatical knowledge (e.g., grammatical knowledge of French) is present in the soul, or color in present in body. I am now in a position to unpack Aristotle s four-fold division of being (τἅ ὄντα). There are things that are said of a subject but not present in a subject. These things are secondary substances. 337 His example of such an entity is man, which is a universal in the category of substance. Next, there are things that are present in a subject but not said of a subject (e.g., this particular knowledge or this particular white). These things are particular or primary attributes that cannot exist independently of their bearers. There is considerable controversy over the exact nature of primary attributes. While there is some agreement that primary attributes are non-substantial individuals, there is major disagreement over whether or not a 336 Categories 1a Categories 2a11-a18.

127 116 primary attribute can occur in more than one individual. For example, is it possible for the particular ruddy color of Dante can recur in some other particular. 338 Next, some things are both said of and present in a subject. These things are nonsubstantial secondary attributes (i.e. non-substantial universals). Examples of these are knowledge (in general) and white (in general, as a type of color). The nonsubstantial secondary attribute (universal) whiteness is said-of many primary substances but is only accidental to them. Lastly, there are things that are neither said of nor present in a subject (e.g., this horse, this dog, etc.). Aristotle calls these primary substances because they are ontologically prior to other substances in two ways: 339 (1) Every secondary substance is said of primary substance (2) Every non-substance (both particular and universal) is present in some primary substance. 338 Ackrill (1963), argues for a strict dependence of non-substantial particulars on their bearers. Frede (1987), and Owen (1965), argue for a more general dependence. Among the things at stake here is the relationship between primary and secondary attributes. As Modrak (2001), 31 points out, this controversy has minimal impact on the claims of Peri Hermeneias 1. All that need be accepted is the view that non-substantial particular attributes are not capable of independent existence. See Wedin (1993), for a review of interpretations regarding this issue. 339 Aristotle argues in support of this claim at Categories 2a34 2b7

128 117 This establishes that every thing that is not a primary substance is either said of or present in a primary substance. Substances are the subjects that ontologically dependent non-substances are in. Each member of a non-substance category therefore stands in this inherence relation to some substance or other. For example, color is always found in bodies and knowledge in the soul. Neither whiteness nor a piece of grammatical knowledge is capable of independent existence. Each requires there be some substance in which it inheres. I can extract the following taxonomy from Categories 1a20-1b7: Table 2 Schema for Aristotle s Categories 340 SUBSTANCE (a) Said of a subject, not present in a subject (e.g., man, dog, animal) NON-SUBSTANCE (c) Said of a subject, present in a subject (e.g., knowledge, white) Secondary Attribute Secondary Substance (d) Not said of a subject, not present in a subject (e.g., This man, this dog) Primary Substance (b) Not said of a subject, present in a subject (e.g., This particular white) Primary Attribute 340 Ackrill (1963), 77.

129 118 In addition to this fundamental inherence relation across categories, Aristotle also points out another fundamental relation that obtains between items within a single category. He describes this as the relation of being said of a subject, and his examples make clear that it is the relation of a more general to a less general thing within a single category. I can further illustrate this type of predication by exploring the categories of substance and quality with the following diagrams: Figure 7

130 119 Figure 8 Each category can be divided into basic kinds. For example, the category of substance can divided into animals, plants, and such. Each of these kinds can be further divided into species such as humans, horses, etc. Lastly, each these genera can also be divided into individuals (e.g., Yoda and Snippy) that are the fundamental members of each category. Thus, Dog is said of a particular dog (e.g., Yoda), and animal is said of dog. Therefore, animal is said of the particular dog also, because the said of relation is transitive. So the genus (e.g., animal) is said of the species (e.g., dog) and both genus and species are said of the particular. The same holds in non-substance categories. In the category of quality, for example, the genus (color) is said of the species (white) and both genus and species are said of the particular white (e.g., the white of Johnny Winter).

131 120 For my purposes the most important connection between Peri Hermeneias and the Categories are said of and present in relations. It is clear from the nesting of attributes above that the primary substances of the Categories are fundamental ontological entities and sensible characteristics (e.g., color, etc.) adhere in them. Perceptual experiences involve sensible particulars (e.g., this particular dog or horse) but the content of our perceptions contains both things said of and present in that particular substance. This content is grasped via the proper, incidental, and common objects of perception. As discussed in section 3.2, perceptual experiences produce φαντάσµατα and the content of those φαντάσµατα become the objects of thought through the activity of the active intellect See page of this thesis.

132 121 CHAPTER FOUR: Signification and Semantic Import Introduction Now that I have explored the nature of the παθήµατα and πράγµατα mentioned in Peri Hermeneias 1, I can turn our focus to a central and controversial concept in Peri Hermeneias 1-6: signification. Traditionally, Aristotle s concept of signification has been viewed as identical to our concept of meaning. In part, this is due to the fact that the etymological root for our term semantic is the Greek word, σηµαίνω (signify), which Aristotle uses primarily in his discussion of signification. 342 So it seems natural to believe that 'signify' means 'means'. As with much of the content of Peri Hermeneias, commentators over time have offered a variety of differing interpretations of the Aristotle s concept of signification. Terrance Irwin and Norman Kretzmann have challenged the traditional account of signification and its semantic import. 343 More recent commentators have argued that Aristotle s concept of signification is a central part of Aristotle s semantic thought. 344 My goals in this chapter are to develop an account of Aristotle s concept of signification and then apply that account to the claims made in Peri Hermeneias Irwin (1982), Irwin (1982) and Kretzmann (1974). 344 Recent entries in the debate have been made by Modrak (2001), Wheeler (1995), and Charles (2000).

133 In section 4.1, I begin my investigation into the semantic import of Aristotle s concept of signification by reconstructing and analyzing the relations between significant sounds, πράγµατα, and παθήµατα found in the schema of Peri Hermeneias 1. Next, in section 4.2, I critically analyze and then rework an account of signification put forth by David Charles. Here, I specifically examine David Charles claim that, in one sense, signification can be understood as a two-place relation between a word and some mental state. He states that, "the significance of a simple linguistic expression 'A' is fixed in the first instance by what the thought with which it is conventionally correlated." 345 I also explore David Charles claim that signification can be thought of, in another sense, as a three-place relation between a word, a thought and a non-linguistic entity in the world. Charles claims that "[I]f we apply the De Anima model in those cases where thoughts are likenesses to things...[the thought conventionally correlated with 'A'] is about the object which causes that thought in the likening (form-transferring mode)." 346 Lastly, in section 4.3, I apply the revised account from section 4.2 to the linguistic phenomena (i.e., names, verbs, etc.) that Aristotle examines in Peri Hermeneias David Charles (2000), Ibid.

134 The relations in Peri Hermeneias 1 As I noted in the previous chapter, there are several relations posited at Per Hermeneias 16a3-9. Following Kretzmann (1974) and Wheeler (1995), I identify nine important claims in the schema of Peri Hermeneias 1: 347 R 1 : Written marks are symbols of spoken sounds R 2 : Spoken sounds are symbols of affections in the soul R 3 : These are signs of affections in the soul 348 R 4 : Each affection of the soul is a likeness of some real thing R 5 : Written marks are not all the same for all humans R 6 : Spoken sounds are not all the same for all humans R 7 : Affections of the soul are the same for all human beings R 8 : Real Things are the same for all humans R 9 : These matters have been discussed in the work on the soul and do not belong to the present subject 347 Kretzmann (1974), 4 and Wheeler (1999), uses the schema A1-A9 to present these relations. 348 The issues and interpretative challenges associated with this claim were discussed by the early Greek commentators and examined in chapter two, section 2.2 of this dissertation.

135 124 These can be organized as follows to highlight the important relations: Table 3 Relations in Peri Hermeneias 16a3-8 Relatum Relation Relatum Affection of the soul is a likeness of Real thing Spoken sound and/or Written Marks is/are a sign(s) of Affection of the soul Spoken sound is a symbol of Affection of the soul Written mark is a symbol of Spoken Sound Two important semiotic relations are posited at Peri Hermeneias 16a The first relation is posited between a significant sound/written mark and an internal state (παθήµα). The second relation posited is between the aforementioned internal state (παθήµα) and an external object (πρᾶγµα). Based on the analysis that I developed in the previous chapter both of these relations can be reconfigured by substituting φάντασµα for παθήµα. The first relation then holds between a significant sound/written mark and some φάντασµα. The second relation holds between a φάντασµα and some πρᾶγµα. Traditionally, these relations are taken to present the basic semiotic relations between mental states, language, and reality. 349 See page of this thesis.

136 125 Commentators from the Neo-Platonist 350 on have claimed that for Aristotle spoken sounds signify in the first place affections of the soul and that by means of this primary significance, spoken sounds signify in the second place real things. 351 I follow Wheeler (1999) in calling this view the Semantic Primacy Thesis (SPT). 352 According to this view spoken sounds/written marks signify immediately affections in the soul and signify mediately things in the world. 353 As a part of the Received View discussed in chapter 3 section 1, 354 this triadic relation between things in the world, mental states, and language has been seen as the heart of Aristotle s semantic theory in Peri Hermeneias. This view has come under attack from several commentators. Kretzmann (1974) claims that the SPT is false. Kretzmann claims that R 1 R 9 do not explicitly nor implicitly establish a relationship between spoken sounds/written marks and extra mental objects. 355 He claims that one has no justification to infer any relationship between spoken sounds/written marks and real things. 356 Kretzmann s rejection of the SPT is based on his rejection of the synonymy of sign and symbol in Peri Hermeneias 350 Both Ammonius and Stephanus hold the view that words are σηµεῖα of παθήµατα πρώτως and σηµεῖα of πράγµατα δευτέρως. See page of this thesis. 351 Wheeler (1999), Ibid. 353 This interpretation is based on reading the adverbial πρώτως at 16a See page 80 of this thesis. 355 Kretzmann (1974), Ibid.

137 126 16a Kretzmann claims that that Aristotle use of πρώτως at 16a6 is evidence that the natural sign relation (of a spoken sound to an affection of the soul) is logically prior to the conventional symbol relation between a spoken sound and an affection of the soul. 358 As I indicated in chapter three, the balance of the philological evidence stands against Kretzmann s claims. 359 Another objection to the SPT centers on interpretive problems, which emerge regarding in the first place at 16a6. As Wheeler (1999) points out there are three major problems that arise. 360 The evidence from the manuscript tradition does not provide a consensus on whether or not πρώτως, πρώτων, or πρῶτον is correct at 16a6. While the majority of the manuscripts contain πρώτως, 361 others have the neuter singular πρῶτον. 362 However, there is evidence that points to 357 See page of this thesis. 358 Kretzmann (1974), See page of this thesis. 360 Wheeler (1999), As Magee (1989), points out the long tradition that defends the SPT as well as Kretzmann (1974) replace πρώτων with πρώτως. Magee points out that Bekker (1831) and Waitz (1844) follow Niphus (1555), Pacius (1566), and Lucio (1619) in making the same interpretive move. Steinthal (1890), Cooke (1938), Colli (1955), and Tricot (1959) also make this move. This interpretation is supported by the occurrence of πρώτως in early MSS (Ambrosianus and Marcianus as well as Urbinas and Coislinianus) as noted by Minio-Paluello (1949). Also Ammonius, Stephanus, and the anonymous Syriac commentator have the adverbial. 362 According to Magee (1989) p24, Montanari (1984) argues that Minio-Palello (1949) failed to notice that πρώτως was a second hand correction of πρῶτον.

138 127 πρώτων being the oldest variant. 363 Second, if one chooses the genitive plural at 16a6 then the passage reads as claim about first things of which significant sounds are signs, but it is unclear whether these first things are first temporally, semantically, causally, etc. 364 Third, if one chooses the adverbial reading at 16a6 then it unclear whether the passage is makes a distinction between the semantic relation of spoken sounds to affections of the soul and that between spoken sounds and real things or whether it differentiates the relation spoken sounds bear to affections in the soul insofar as spoken sounds are natural signs as opposed to the relation spoken sounds bear to affections in the soul insofar as spoken sounds are conventional symbols. 365 Given the mixed philological evidence for the SPT, I follow Wheeler (1999) in focusing on other grounds to establish that natural relation of spoken sounds to affections in the soul is semantically prior to the conventional relation between spoken sounds and affections in the soul. 366 As I have shown in the previous chapter, every spoken sound is a vocal 367 sound and every vocal sound in Montanari s claim is supported by physical evidence of the erasure of the last three characters of the word and the existence of a circumflex accent in the original hand. 363 Magee (1989), argues that the oldest translation of Peri Hermeneias including all the Boethian manuscripts, the ancient Armenian version, Vaticanus Barberianus, and Georgianus have the genitive plural. 364 Wheeler (1999), Wheeler (1999), Ibid. 367 See page of this thesis.

139 128 accordance with its nature is accompanied by an act of φαντασία that is its σηµαντικός (signification). 368 Therefore, every vocal sound by nature signifies some φάντασµα. This not only provides a strong naturalistic component to Aristotle s account of signification, but also supports the semantic priority of the relation of spoken sounds to φαντάσµατα in the soul. All vocal sounds are natural signs of φαντάσµατα in the soul, but are also potentially conventional signs of φαντάσµατα in the soul. Although the range of vocal sounds is the same for all humans, those adopted by different linguistic communities may differ. Having found grounds for the SPT that are not susceptible to the major objections, I am now in a position to examine the mechanics of the likeness relation and the account of signification outlined in Peri Hermeneias 16a3-9. In the next section, I critically examine a recent compelling interpretation by David Charles that holds some promise for clarifying these relations. 368 See page 89 of this thesis.

140 Likeness and Signification Now that I have established that there are strong philosophical and philological grounds for attributing to Aristotle the semantic priority of natural relation of spoken sounds to affections in the soul is over the conventional relation between spoken sounds and affections in the soul, I can explore in depth the relation between φαντάσµατα and πράγµατα. As I discussed in the previous chapter, Aristotle s indicates that the relationship between φαντάσµατα and πράγµατα in the work is one of likeness. 369 As pointed out in the previous chapter, the implications of this are that φαντάσµατα resemble real things in that φαντάσµατα must have some properties (qualities) in common with πράγµατα Likeness Charles (2000) presents an interpretation of Aristotle s semantics for simple names according to which a speaker uses a name to signify iff she does so by means of a thought causally connected to the significate of that name. 371 Charles (2000) claims that successful thinking, like perception, involves entities in the world 369 See page of this thesis. 370 Ibid. 371 Charles (2000),

141 130 liken[ing] the relevant faculty to themselves. 372 The likening process for thoughts requires: A properly functioning thinking faculty; Some external object as its starting point; The external object to be the efficient cause of the thought 4. [T]he patient [to be] made like the agent in certain relevant respects. 376 On this view, likening is an asymmetric process where the relevant object or kind is the efficient cause of the thought and explains the general features of that thought (i.e. its content, etc.). 377 The role of the efficient cause is central to Charles s account. It is a particular account of this role that I use to develop my account of likeness and signification in the next section. I begin by examining the key parts of Charles likeness account numbered (1) (4) above. I begin with claims (1) and (2) and finish my analysis with claim (1). As I discussed in Section 3.2, the starting point for thought just as the starting 372 Charles (2000), Charles (2000) discusses the likening process at two places and Charles (2000), Charles (2000), Ibid. 377 Charles (2000), 82.

142 131 point for perception is some external object. 378 Just as perception involves the reception of sensible form from some external object, thinking involves the reception of some intelligible form that inheres in an external particular. 379 The intelligible form is transmitted via φαντάσµατα that are produced as a result of the perceptive experience. 380 As I discussed earlier, thinking involves the noetic capacity becoming like its object. This can only be brought about when the thinking faculty is functioning properly To explain claims (3) and (4), Charles points to De Generatione Corruptione 324a5-7 where Aristotle claims that if one thing affects another then they must be distinct in species but the same in genus. 382 Initially, the agent and the patient are similar and dissimilar. They are alike enough that the agent is able to activate the patient s potentiality for being thus affected. For example, the hot thing and the cold thing are dissimilar in that one is hot and one is cold. However, the hot thing is similar enough to the cold thing for the hot thing to heat the cold thing. The thing that is potentially hot becomes actually hot when the actually hot thing modifies the 378 See page of this thesis. Also see Charles (2000), ; Modrak (2001), , and Modrak (1987), See page 91 of this thesis. 380 See page of this thesis. 381 Charles (2000), Charles (2000), 114, he points out that the patient and the agent are distinct in form but with the same type of matter.

143 132 cold thing and makes the cold thing like itself. 383 The result of this is that the patient is likened to the agent s active power. As Charles (2000) points out, this process is not one that results in a mere resemblance in terms of quality. 384 Instead this process emphasizes the central role of efficient cause. One of the benefits of this account is that it rules out cases of accidental similarity as cases of likeness. Because the causal role of the agent is central, two objects that have the same degree of some quality (e.g., heat) without any causal connection would not be considered to be likenesses. Support for this account can be found in the Metaphysics where the efficient cause plays a central role in the coming to be of states or things. 385 This model is also central to Aristotle s account of perception. At De Anima 424a17, Aristotle compares the senses reception of the sensible forms without matter with wax receiving the design but not the matter of the signet ring. The wax does not merely resemble the design of the signet ring. The wax becomes a likeness of the signet ring in that it is causally affected by the ring and made like the ring in relevant ways. Further, as Charles (2000) suggests, the imprint [of the ring] is an imprint of the type it is because it is printed in by a ring of this type. 386 Lastly, support for this account can be found in Aristotle choice of παθήµατα for the likeness state. As Modrak (2001) indicates, [t]he mat suffix indicates a state that 383 De Generatione Corruptione 323b29 324a Charles (2000), a and 1050a Op. Cit., 115.

144 133 is the result or effect of an action. 387 The action here is the efficient causation performed by the agent upon the patient. Importantly, this model can be seen at work within the functioning of the noetic capacity. As I discussed in chapter 3.2, 388 the thinking faculty while impassible [is] capable of receiving the form of an object. 389 The noetic capacity is potentially identical with its object. 390 As I pointed out earlier, in De Anima III.5 the active intellect is described as it is described as maker of all things. 391 An analogy is made between the relationship between the active and passive intellect and the relationship between an art or skill and its material. 392 This analogy is significant because skill is presented as the paradigm efficient cause in Physics II As I indicated in Section 3.2, the key to understanding the causal role of the active intellect is the analogy made between the active intellect and light. 394 Just as light makes potential colors actual through its activity, the active intellect makes universals actual through its activity. 395 As such, the active intellect is the efficient 387 Op. Cit., 256 and Smyth (1920) section See page 91 of this thesis. 389 De Anima 429a De Anima 429a De Anima 430a12 and 430a De Anima 430a Physics 195b De Anima 429a3-5 and 430a See page of this thesis.

145 134 cause of thoughts in the passive intellect Signification Now that I have explored the model for likeness presented by Charles (2000), I can connect this model to the discussions in the previous chapters to form an account of Aristotelian signification. I noted at the beginning of chapter 3 and chapter 4 that Aristotle presents several key relations in Peri Hermeneias 16a There is the relation between a spoken sound/written mark and a παθήµατα in the soul and the relationship between this παθήµατα and some external object. Traditionally accounts of signification have focused on one or both of these relations. In what follows I argue that both of these relations have a vital role in Aristotle s concept of signification. I begin by exploring the first of these relations (i.e. the relation between spoken sounds/written marks and παθήµατα). At 16a1-2, Aristotle claims that written marks are symbols of spoken sounds and spoken sounds are signs of παθήµατα. When I apply the account I have developed to the constituents of this relation (i.e. spoken sounds and παθήµατα), it is clear this first relation holds between a spoken sound /written mark and some φάντασµα. As I showed in Section 3.2, the relation thus depicted coheres with Aristotle s biological account of 396 De Anima 430a See page of this thesis.

146 135 language. 398 As I pointed out, Aristotle states that every spoken sound is a vocal sound. Specifically, every spoken sound is an articulate vocal sound and thus must be accompanied by an act of linguistic imagination that is its content or signification. 399 Therefore, signification involves the relation between articulate sounds and some φάντασµα. The strong conventional and natural elements in this relation are noteworthy. Spoken sounds and written marks are conventional signs that are not the same for all. 400 As Wheeler (1999) indicates spoken sounds qua spoken sounds are conventionally significant in virtue of some linguistic community adopting this or that set of letters and syllables for some [παθήµατα]. 401 This establishes a strong conventionalist element in Aristotle semantic thought. Yet it is evident from Aristotle s biological account of language that spoken sounds qua vocal sounds are by nature signs of φαντάσµατα. This establishes a strong naturalist element in Aristotle s semantic thought. It is important to note that the natural element in this account is prior semantically to the conventional element. 402 In the next section, I demonstrate that this account of this relation is consistent with what Aristotle says in Peri Hermeneias 1-6. Let us now move on to the second relation (i.e. the relation between a παθήµατα and some external object). At 16a7-8, Aristotle claims that the παθήµατα 398 See page of this thesis. 399 See page of this thesis. 400 Peri Hermeneias 16a Wheeler (1999), See page of this thesis.

147 136 are likeness of actual things. This relation can be seen to hold between a φάντασµα and some real object. As I pointed out in sections 3.2 and 4.2.1, the likeness relation holds between some φάντασµα and some real thing. This relationship is based upon the likening model in perception where the sensory organ becomes like its object. As I indicated in Section 3.2, the human soul is so constituted that its perceptual organs take on sensible characteristics of their objects as part of perception. 403 The perceptible object affects the sense by en-forming it. 404 Φαντάσµατα are traces of these perceptions that are stored in the sense organs. 405 By their nature, φαντάσµατα are likenesses of real things. Unlike the relation between a spoken sound and a φάντασµα, the relation between a φάντασµα and a real thing does not contain any conventional elements. This explains Aristotle s claim at 16a6 that affections of the soul (i.e. φαντάσµατα) are the same for all. There has been controversy over how to properly understand the role of these relations in Aristotle s concept of signification. Traditionally under the SPT, signification is equated with meaning, and these relations have been seen as committing Aristotle to the view that spoken sounds signify in the first place νοήµατα and secondarily signify things in the world. 406 In opposition to this tradition, Kretzmann (1974) argued that signification is concerned with connection 403 See page of this thesis. 404 Ibid. 405 See page 99 of this thesis. 406 See page of this thesis.

148 137 between mental images and spoken sounds. 407 He also claims Aristotle does not posit a relation between real things and spoken sounds. 408 Irwin (1982) also opposes the traditional view of signification as well. He claims that Aristotle is not concerned with meaning 409 but with various types of signifiers and essences in the world. 410 Most recently Charles (2000) has argued that a spoken sound signifies in the first instance the thought with which it is conventionally correlated. 411 In the second instance, Charles (2000) claims that a spoken sound signifies some kind or object in the world by means of a thought produced in the form-transferring route. 412 On Charles account signification is a semantic account. It causally connects language, our thinking faculties, and a world of essences. On this account, a speaker can use a word to refer to a kind in the world because her thought is causally connected to the kind in the world. For example, when she grasps that thunder is a certain kind of noise in the clouds. This allows the thinker to grasp the non-accidental features of a given kind. While I agree with Charles (2000) that signification is a flexible concept and Irwin (1982) that signification concerns essences in the world, I do not agree with their respective construal of signification. 407 Kretzmann (1974), Ibid. 409 Irwin (1982), Irwin (1982), Op. Cit., 82-83; this is what Charles calls immediate signification. 412 Op. Cit., 83; this is what Charles calls immediate signification. The form transferring route was discussed in section of this dissertation section 4.1.

149 138 On the interpretation I have developed, signification is in one sense a relation between spoken sounds and some φάντασµα and in another sense is a relation between a spoken sound and a thing in the world by means of a φάντασµα. Taken together these relations have important implications for Aristotle s conception of signification. It follows from these two relations that every spoken sounds is a sign of some sensible or intelligible characteristic of a φάντασµα and each φάντασµα is a likeness of some perceptible and/or intelligible quality possessed by some real thing. As I have argued φαντάσµατα are the vehicles that carry content from the perceptual faculty to the noetic faculty. 413 In order to accomplish this, a φάντασµα must be able to represent content that contains sensible and intelligible characteristics. It follows that φαντάσµατα have as their content perceptive experiences of sensible particulars. These sensible particulars are tokens of many types. As I indicated in section 3.3, the content of our perceptions of these sensible particulars contains things said of and/or present in that particular substance. 414 This content grasped via the proper, incidental, and common objects of perception is the content of φαντάσµατα. The result of this is that the noetic faculty takes on the intelligible characteristic(s) contained in the content of a φάντασµα. This makes it possible for φαντάσµατα to represent all types of phenomena including experiences, people, events, etc. In addition, φαντάσµατα play a central role in cognition. Aristotle claims they are to the thinking faculty what 413 See page of this thesis. 414 See page of this thesis.

150 139 perceptible forms are to the perceptive faculties and the soul cannot think without them. 415 As I have indicated the likening relation for thought involves the passive intellect becoming like the object of thought through the activity of the active intellect. 416 Just as light makes the potentially visible actually visible through its activity, the active intellect makes the potentially intelligible content within a φάντασµα actual. 417 Since φαντάσµατα are a key component of signification, the concept must be flexible enough to deal with the robust sensory and intelligible content of φαντάσµατα. As such signification is equipped to deal with both denotational and ideational aspects of meaning. On denotational theories of meaning, the meaning of a term is what the term denotes or refers to. Every spoken sound denotes some sensory or intelligible characteristic possessed by some particular thing in the world. Therefore, reference to the world is secured by the causal connection between φαντάσµατα and things in the world. The φαντάσµατα are a likeness of some real objects in the world and their properties. On ideational theories of meaning, the meaning of a term is some cognitive content (e.g., idea, concept, thought, etc.) correlated with the term and present in the mind of someone who knows and understand it. 418 On the interpretation that I have 415 Aristotle makes these claims in at least three places in De Anima: 431a14-16; 431b6-8; 432a See page of this thesis. 417 Ibid. 418 Wheeler (1999), 214.

151 140 presented, every spoken sound signifies some φάντασµα. The content of that φάντασµα (sensory or intelligible) is in the mind of competent speakers. This account of signification coheres with Aristotle s account of definition in the Posterior Analytics. Irwin (1982) states that names and their definitions signify essences. Irwin implies that there are two species of signification: conventional signification (to us) and natural signification (by nature). 419 What names signify to us are nominal essences and what they signify by nature are real essences in the world. 420 In formulating his sciences, Aristotle is concerned with moving from the former to the latter. I believe that Irwin is correct here and signification, as I have presented it, can accommodate both of these conceptions. In one sense, what a spoken sound signifies is the sensory or intelligible content of a φάντασµα. It follows that a term like dog would pick out and have as its content a nominal essence based upon the sensory content of φαντάσµατα caused by perceptual experiences of dogs in the world. A speaker who uses a name in this way need not know that there is some object that was the efficient cause of the φάντασµα. This is the type of signification at work in Posterior Analytics II.8, where Aristotle states that one type of definition is what a name or a name-like phrase signifies. 421 These types of nominal definitions do not require that the referent of the name (or name-like expression) exist. In another sense, what a spoken sound signifies is the 419 Irwin (1982), Irwin (1982), Posterior Analytics 93b30-32.

152 141 intelligible content of a φάντασµα. In this way, a term like dog would pick out and have as it content a real essence abstracted from the content of φαντάσµατα by the active intellect. 422 This will accommodate another type of definition in Posterior Analytics: real definitions. A real definition states the essential nature of a thing and explains a thing s existence See p of this thesis. 423 Posterior Analytics 93b29-94a11.

153 Signification and Peri Hermeneias 1-6 Having developed an account of signification, I am now in a position to explore Aristotle thought in Peri Hermeneias 1-6. I do so by intrepreting Aristotle s account in light of the interpretation I have developed. I focus on three main linguistic phenomena discussed by Aristotle. At the beginning of Peri Hermeneias Aristotle states, First we must settle what a name [ὄνοµα] is and what a verb [ῥῆµα] is, and then what a negation [ἀπόφασις], an affirmation [κατάφασις], a statement [ἀπόφανσις], and a sentence [λόγος] are. 424 Over the course of the first six chapters Aristotle makes important claims regarding their signification. In this section, I examine Aristotle s claims regarding each of these linguistic items in turn Ὄνοµατα (Names) At 16a Aristotle states, A name (noun) [ὄνοµα] is a spoken sound [φωνὴ] significant by convention [σηµσντικὴ κατὰ συνθήκην], without time [ἄνευ χρόνου], none of whose parts is significant in separation [ἧς µηδὲν µέρος ἐστὶ σηµαντικὸν κεχωρισηένον]. For in 'Whitfield' ( Goodsteed ) [Kάλλιππος] the 'field' ( steed ) [ἵππος] does not signify anything in its own right [ουδὲν αὺτο καθ ἑαυτὸ σηµαίνει], as it does in the phrase 'white field' [καλὸς ἵππος] ( good steed ). Not that it is the same with complex names [συµπεπλεθµένοις] as with simple ones a1ff.

154 143 [ἁπλοῖς ὀνόµασιν] 425 : in the latter the part is in no way significant [τὸ µέρος οὐδαµῶς σηµαντικόν], in the former it has some force (want to/pretend to signify) [βούλεται] but it is not significant [οὐδὲν σηµαίνει] of anything in separation, for example the 'boat' in 'pirate-boat' [ἀλλ οὐδενος κεχωρισµένον, οἶον ἐν τῷ ἐπακτροκέλης τὸ kέλης οὐδὲν σηµαίνει καθ ἐαυτό]. When I apply the account that I have developed in the preceding chapters, it can be seen that names are vocal sounds that are conventional signs for φαντάσµατα in the soul. The conventional connection between the name and the φάντασµα is reinforced at 16a27 where Aristotle states, I say by convention [κατὰ συνθήκην] because no name is a name naturally (no sound is by nature a noun) [φύσει] but only when it has become a symbol [σύµβολον]. 426 A particular set of phonemes becomes a sign/symbol of a φάντασµα (caused by some object in the world) when it is adopted as such by a linguistic community. This however presents a quandary. Are names in this context to be taken as signs of the sensory content of φαντάσµατα or the intelligible content of φαντάσµατα that is abstracted by the active intellect? I argue that the correct answer for Aristotle here is that names ought to be taken as signs of the intelligible content of φαντάσµατα. While at this point I cannot completely rule out nominal accounts of the significance of names, there are good 425 Cooke (1938) translates this, as It is necessary to notice, however, that simple nouns differ from composite. 426 cf. Cratylus 386e-390e (discussed on page of this thesis).

155 144 reasons to believe that Aristotle focuses in Peri Hermeneias on intelligible content of φαντάσµατα. If names are taken here to be merely signs of sensory content of φαντάσµατα, then the content of names would at best be limited to nominal essences drawn from the sensible features of particulars. There are good reasons to believe that this is not what Aristotle intends in Peri Hermeneias and other parts of the Organon. 427 As I indicated in the previous section of this chapter, real definitions are an essential part of Aristotelian science. I show in this section that Aristotle has in mind these types of accounts when it comes to a specific class of statements that can be evaluated for truth and falsity. Aristotle states that the parts of names do not signify in separation. This separates his work from the mimetic account in the Cratylus that I examined. 428 For Aristotle neither names nor their parts are likeness of their nominata. On the account of language that I have developed in the previous chapters, there is no necessary connection between one or another sets of phonemes and the content of some φαντάσµατα. 429 The issue of whether or not the parts of names are significant is a particular puzzle for complex names. Aristotle claims that the parts of complex names have some force but do not signify separate things when combined in a single name. On 427 The distinction between real and nominal definitions is a central concern in the account of definition in Posterior Analytics II This is discussed on p of this thesis. On Cratylus account the individual phonemes that name up the name are likenesses of the nature the nominatum. 429 This was discussed in section of this thesis.

156 145 first blush it would appear that the parts of a complex name like pirate-boat should signify when separated. For, both pirate and boat are legitimate names in their own right. There seems to be no reason to that either of these names could not be signs of some φαντάσµατα in their own right. If asked Aristotle s answer might well be that the φάντασµα conventionally connected to pirate-boat is not the same as the one conventionally connected to pirate or boat. The force that these words would seem to have comes from the similarity these names have to the names correlated with the φάντασµα associated with boat. The boat from pirate-boat sounds the same as the simple name boat but they are associated with different φαντάσµατα. The φάντασµα correlated with the simple name boat is the sensory or intelligible content that arises from perceptual experiences of real object(s) (i.e. boats). The φάντασµα correlated with the complex name pirate-boat is a different set of sensory or intelligible content that arises from perceptual experiences of some different real object(s) (i.e. pirate-boats). Since not every boat is a pirate-boat, the objects that are causally connected to the respective φαντάσµατα of boats and pirate-boats are not the same and it follows that the content of the respective φαντάσµατα are not the same.

157 Ῥῆµατα (Verbs) Aristotle states that a verb is in some ways similar to a name. Verbs are significant sounds whose parts are not significant in separation. 430 In addition, he indicates that a verb, when uttered by itself, is a name [ὀνόµατά] (noun) and signifies [σηµαίνει] something (stands for something). 431 As such, a verb is a vocal sound conventionally correlated with the content of some φάντασµα. However, Aristotle indicates a verb is also different from a name in that it additionally signifies time. 432 To illustrate this difference, Aristotle points out that, 'recovery' ( health ) [ὑγιεια] is a name (noun) [ὄνοµα], but 'recovers' ( to be healthy ) [ὑγιαίνει] is a verb [ῥῆµα], because it additionally signifies something's holding now [προσσηµαίνει γαρ τὸ νῦν ὑπάρχειν]. 433 Aristotle indicates that φαντάσµατα has a role in our awareness of the passage of time. 434 The perception of time involves the use of memory to recall the perception of nows other than the present. 435 Φαντάσµατα are the 430 Peri Hermeneias 16b Peri Hermeneias 16b Peri Hermeneias 16b Cooke (1938) translates this last phrase as but also conveys that the state signified (namely, health) now exists. 434 De Memoria 452b15-22 and 452b Physics 220a18-19, and Also, De Memoria 450b20 451a16.

158 147 vehicles by which one can recall a diverse class of things including time. 436 As Aristotle points out in the Physics, we do not perceive time καθ αυτά but κατά συµβεβηκός. 437 It follows that the content of the φαντάσµατα correlated with a verb involves the common and incidental objects of a φαντασία. This account also holds for inflexions of verbs, because they are distinguished from verbs due to the fact that inflexions signify time outside of the present. 438 Aristotle also says a verb is always a sign (indication) [σηµεῖον] of what holds, that is, holds of a subject (of the things belonging to a subject or in a subject) [οἷον τῶν καθ ὑποκειµέµου ἢ ἐν ὑποκειµένῳ]. 439 This reveals the particular content of a φάντασµα with which a verb is correlated. A verb is a predicate and as such is correlated to a sensible attribute that is said of and/or present in some sensible particular. The content of the φάντασµα 436 In De Memoria, Aristotle lists a number of items that can be remembered. At 449b20-2, he indicates that one can remember facts (i.e. something seen, heard, contemplated or learned). At 452a14-16, he claims that one can remember the last thing in a list such as milk, white, moist, fluid, and autumn. At 451a29, he states one can remember objects of scientific knowledge. At 452b17-453a1, he indicates that one can remember events such as something one did two days ago. One can also remember a person (45b30-451a1) or a name (452b5) or the past (449a15). 437 Physics 219a Peri Hermeneias 16b Cooke (1938) translates the preceding phrase as an indication of something asserted of something; I mean, of a something predicated of a subject or found present in it.

159 148 correlated with a verb contains the sensible and intelligible characteristics that arise from perceptual experiences of some sensible particular Indefinite names and verbs Indefinite names Much has been written concerning Aristotle s account of indefinite names and indefinite verbs. 440 However, Aristotle s own comments regarding them are limited to a few lines in Peri Hermeneias. At 16a30-31 Aristotle claims, 'Not-man' [οὐκ ἄνθρωπος] is not a name (and the like are not nouns), nor is there any correct name (established/set down name) [κεῖται ὄνοµα] for it (I know of no recognized names we can give such expressions). It is neither a phrase [λόγος] nor a negation [ἀπόφασις] (which are neither denials nor sentences) [οὐ µὴν οὐδὲ κεῖται ὄνοµα ὅ τι δεῖ καλεῖν αὐτό]. Let us call it an indefinite name (indefinite nouns) [ὄνοµα ἀόριστον] (since we use them of all kinds of things, non-existent as well as existing) [ὅτι ὁµοίωσ ἐφ ὁτουοῦν ὑπάρχει καὶ ὄντος καὶ µὴ ὄντος]. At 19b6-10 he restates that 'not-man' is an indefinite name, but also claims that indefinite names signify in a way one thing but something indefinite. Regarding 440 See Irwin (1982), Bäck (2000), Wheeler (1999), Modrak (2001), Charles (2000) for a discussion of indefinite verbs.

160 149 indefinite verbs Aristotle states at 16b11-18, 'does not recover' ( Is notwell ) [οὐχ ὑγιαίνει] and 'does not ail' ( is not-ill ) [οὐ κάµνει] I do not call verbs [οὐ ῥῆµα λέγω]. For though they additionally signify time and always hold of something, yet there is a difference - for which there is no name [προσσηµαίνει µὲν γὰρ χρόνον καὶ ἀεί κατά τινος ὑπάρχει, τῇ δέ διαφορᾷ ὄνοµα οὐ κεῖται]. 441 Let us call them indefinite verbs [ἀόριστον ῥῆµα], because they hold indefinitely of anything whether existent or non-existent [ὅτι ὁµοίως ἐφ ὁτουοῦν ὑπάρχει, καὶ ὄντος καὶ µὴ ὄντος]. In addition, Aristotle hints at 19b10 that indefinite verbs also signify in a way one thing but indefinite. 442 Irwin (1982) claims that indefinite names like Not-Man signify something indefinite because they fail to signify a single nature. 443 Surely, Irwin seems to be correct. Some things that are not-man include ducks, cats, bees, chairs, angels, hobbits, and lectroids. Clearly, this is a heterogeneous class of things. However, on the account that I have presented, indefinite names are indefinite either because they are not causally correlated to the sensory content of a φάντασµα or they are not causally correlated to some intelligible essence contained in a φάντασµα. 441 Cooke (1938) translates this as Though they certainly have the time-reference and function at all times as predicates, I know of no recognized name. 442 This claim about indefinite names is extended by Aristotle to claims about indefinite verbs about indefinite verbs such as does not recover. 443 Op. Cit., 244.

161 150 While something similar holds true for indefinite verbs, they additionally are indefinite because they belong alike to anything that is and that is not. 444 In other words they are indefinite because they do not pick out a specific attribute contained within the sensory or intelligible content of a φάντασµα λόγος (sentence) Aristotle claims, A sentence [λόγος] is a significant spoken sound (significant speech) [φωνὴ σηµαντικὴ] some part of which is significant (have meaning) [σηµαντικόν] in separation. 445 Aristotle indicates that, Every sentence [λόγος] is significant (not as a tool (instrument of nature) [ὄργανον] 446 but, as we said, by convention [κατὰ συνθήκην]). 447 Aristotle points out that his focus in Peri Hermeneias is on sentences in which there is truth or falsity [ἐν ᾧ τὸ ἀληθεύειν ἤ ψεύδεσθαι ὑπάρχει]. 448 These types of sentences are statement-making sentences [λόγος ἀποφαντικὸς]. He also claims every statement-making sentence [λόγος ἀποφαντικὸς] must [ἀνάγκη] contain a verb [ῥήµατος] or an inflexion of a verb 444 Peri Hermeneias 16b15 and Irwin (1982) p Peri Hermeneias 16b Cooke (1938) translates οργανον as instrument of nature. 447 Peri Hermeneias 17a1-2; this has been taken by commentators from the Neo- Platonists on to be a reference to the naturalist position espoused by Cratylus in the Cratylus. 448 Peri Hermeneias 17a2-3.

162 151 [πτώσεως ῥήµατος]. 449 Aristotle states that there are two main types of statementmaking sentences: affirmation [κατάφασις] and negation [ἀπόφασις]. 450 Sentences: Figure 5 I begin by recalling my previous diagram of Aristotelian Statement-Making An affirmation [κατάφασις] is a statement [ἀπόφανσίς] affirming something of something [τινὸς κατά τινός], a negation [ἀπόφασις] is a statement [ἀπόφανσίς] denying something of something [τινὸς ἀπό τινός]. 451 Based on the account that I have developed the name in an affirmation or negation signifies either the sensory content of a φάντασµα or the intelligible content of a φάντασµα. The verb or inflexion of a verb in an affirmation is some attribute said of and/or present in some sensible particular and contained in the content of a φάντασµα. 449 Peri Hermeneias 17a Peri Hermeneias 17a Peri Hermeneias 17a25-26.