BMCR 2006.05.17

Propositional Perception: Phantasia, Predication and Sign in Plato, Aristotle and the Stoics

, Odysseus, hero of practical intelligence : deliberation and signs in Homer's Odyssey. Lanham, MD: University Press of America, 2004. ix, 377 pages ; 22 cm. ISBN 076183026X $50.50 (pb).

This book offers an analysis of the philosophical development of ancient Greek theories of perception ( phantasia) and of the sign ( sêmeion, tekmêrion), from Plato and Aristotle to the Stoics, Epicureans, and Sceptics.1 In general, it is with regard to his examination of the relationship between these two notions that Barnouw’s study is most original and intriguing. But prevalent flaws of exposition (e.g., unclear and overly technical writing, incoherent organization, lack of explicit argumentation, sloppy editing) make it difficult to follow the author’s discussion and ultimately to evaluate its success and potential usefulness. The most fitting audience for the book would be scholars who specialize in Stoic logic and its philosophical contexts.

The author’s ultimate concern in this book is Stoic logic and its appeal to the notions of perception and sign, especially as the Stoics inherited and developed these notions within the context of the views of Plato and Aristotle.2 Barnouw’s overall thesis is that the Stoics understood phantasia (in their sense, a mental image that contains an intrinsic claim about reality) to be the foundation of knowledge in that it “underlies” a continuity between perception, predication, sign-inference and demonstrative proof.3 The Stoics’ theory, in turn, brought about a “fundamental recentering of epistemology in perception and experience” (4), such that they made important steps towards empiricism and the semeiotic logic of C. S. Peirce. The fact that Stoic logic is, as it were, the telos of this book becomes especially evident in the somewhat distracting proleptic tendency of the author’s early discussions of Plato and Aristotle, which he regularly interrupts with comparisons to the Stoics and with conclusions that he has yet to establish (e.g., pp. 10-13, 16, 19, 25-26, 34, 43, 98, 125, 142, etc.).

The book begins with an exceedingly abstract introduction that does not do a very good job of orienting the reader to the general topic, its ancient contexts, relevant texts, the author’s method and terminology, and previous scholarship. A clear and systematic introduction would have helped considerably to make the author’s arguments and conclusions more accessible. In chapter one, Barnouw examines Plato’s notions of phantasia and predication on the grounds that they set the terms for Stoic logic. He argues that Plato in the Theaetetus coins the term phantasia as ‘misleading appearance’ in order to criticize the tendency of taking the senses as revealing reality. In turn, Plato suggests in the Sophist that the tendency of phantasiai to mislead is due to their capacity to be true or false. Barnouw will go on to argue that while the Stoics adopt Plato’s conception of phantasia as a combination of appearance and assertion (i.e., “propositional perception”), at the same time they reverse this conception by attempting to establish confidence in what phantasiai reveal.

The second chapter considers Aristotle’s definition of phantasia in de Anima III.3. Barnouw argues that Aristotle attempts to correct Plato’s suggestion that phantasia is a misleading form of judgment through perception. Instead, Aristotle understands it to be the representational aspect of perception and judgment, itself conveying no intrinsic claim about reality. Barnouw also shows that although Aristotle tries to free phantasia from its involvement with judgment and the consequent derogatory implications, he does not, however, remain consistent in that attempt, since in certain contexts he speaks of phantasiai as being true or false.

Chapter three examines Aristotle on the relationship between perception, predication, and sign. The author argues that Aristotle’s focus on demonstration to the exclusion of induction created a “gap” in his epistemology, such that he neglected to identify a crucial connection of perception to proposition and proof. The Stoics, Barnouw maintains, would later fill this gap with their “more highly integrated” approach to logic (97-99). Aristotle himself could have avoided the deficiency if he had developed more completely his theory of the sign and had connected it to his analysis of perception. Nevertheless, Barnouw argues, his treatment of the sign marked a significant advance and proved to be indispensable to Stoic semiotics. The author concludes that Aristotle might even have developed a theory of empirical knowledge if he had taken the sign more seriously. He takes this stance against Myles Burnyeat’s view that in his account of signs Aristotle was the first to recognize that there do exist compelling, and yet formally invalid, arguments, and thus was the first to acknowledge the important connection between signs and human rationality.4

In the three remaining chapters Barnouw goes on to examine the Stoic theory of the sign and its relation to phantasia, proposition, and proof. Ultimately he intends to show that the Stoics, and not Aristotle, were the first to make perception and sign fundamental to human thought, thereby taking an important step in a transition towards empiricism in epistemology and logic. The author argues in chapter four that the Stoic notion of phantasia does not refer to impression, appearance, or imagination, but rather means ‘perception’, that is, a mental image that includes an intrinsic assertion about reality. Stoic phantasiai, therefore, say something about what is going on in the world. Their objects are events or states of affairs, as they both establish a relation between some subject and an attribute and assert the reality of that relation. Barnouw goes on to argue that Chrysippus connected this propositional notion of perception to proof through his inferential conception of the sign. He explains:

[Chrysippus’] conception of the sign as a type of inference (‘if this, then that’) puts into technical logical form a means of thought which the Stoics see as fundamental to life and important to the practical arts…. Emphasis on this sign informs a Stoic transformation of logic, focusing on the conditional, as opposed to the categorical (as in Aristotle), and on relations between propositions rather than between terms or classes…. The shift from ‘all x are y‘ to ‘if x, then y‘ gains in significance when the variables… are not terms but propositions. This accords with the Stoic approach to reality as causally determined series of events, not … a hierarchy of entities defined by essences and classes (154).

The fifth chapter examines the Stoic conception of the indicative sign within the polemic context of Sextus Empiricus’ writings. Barnouw argues that Sextus’ account of Stoic semiotics is unreliable. First, it is the Stoics’ strict logical definition of the sign to which Sextus appeals in his attempt to contradict their remarks about the less conclusive signs of everyday life. Secondly, Sextus’ distinction between ‘indicative’ and ‘commemorative’ signs (terms that Barnouw renders, respectively, ‘revealing’ and ‘recalling’) neglects to accommodate the Stoics’ indicative signs because it wrongly restricts their significata to items that are non-evident by nature (that is to say, items that cannot appear alongside their signs). Instead, Barnouw defends a less-restrictive definition of the Stoics’ indicative sign (i.e., a sign that has not appeared alongside its significatum). This construal, he contends, does remain consistent with their strict logical definition. The author finds additional support for his view in Galen’s “more balanced” account of the indicative/commemorative distinction in the debates between the Hellenistic medical schools, where, as Barnouw argues, it seems to have originated. He concludes that Sextus’ criticism ultimately backfires because it shows that the Stoics recognized the continuity between life and logic and maintained an interest in signs throughout a wide range of human activities and degrees of validity.

In the sixth and final chapter the author discusses the Stoics’ strict logical definition of the sign and its relationship to their notion of demonstrative proof. Barnouw emphasizes that Stoic logic deals primarily with the truth of fact, events or situations, as opposed to pure logical relations without reference to reality; he thus characterizes it as “not formalist.” He defends this view against the position of ancient Sceptics and many modern scholars that the Stoics’ strict logical conception of the sign precludes any concern on their part with everyday inferences and probable knowledge. Given its empiricist characteristics, Barnouw notes that it is ironic that the recent revival of interest in Stoic logic derives from the rise of formal logic and from attempts to assimilate Stoic doctrine to Frege’s system, despite the fact that it more precisely anticipates the semeiotic logic of Peirce. Barnouw concludes with an account of how the Stoics go about converting signs drawn from experience into logically conclusive signs.

The author appends to his main discussion two essays of moderate length. The first offers a revision to Charles Kahn’s analysis of the senses of the Greek verb ‘be’ so as to defend a broader, and thus more accurate, conception of classical predication as being “assertoric” of reality. The second appendix examines the relations between Stoic logic and that of Peirce, who himself recognized the Epicureans, and not the Stoics, as his own precursor.

Barnouw’s book has several commendable qualities. As mentioned above, the idea is original, interesting and promising. His discussions of ancient philosophical texts and the issues that they raise are clearly very knowledgeable, and his philosophical analyses rigorous and detailed. All things considered, the book likely will be welcomed by those who are at home with the texts and issues involved in the study of Stoic logic and epistemology. But to the extent that good philosophy is clear and explicit writing, Barnouw’s book leaves much to be desired (see the Appendix below for examples of these problems). One consequence of this shortcoming is that, having finished the book, the reader is left wondering what exactly Barnouw means by the “link” and “continuity” between perception, predication, sign, and proof. What is the nature of this continuity and exactly how is it “carried over” (317) into proposition, sign and proof? What precisely is it about Stoic perception and sign-inference that allows them to “ground” (278) the conclusive power of logical argumentation? These are just some examples of the writer’s tendency to leave much implicit, thus forcing the reader to take extra effort to piece together his meaning. Again, formal conclusions to each chapter that state explicitly the status of the argument would have helped tremendously.


This appendix surveys some examples of the flaws of writing and editing that make this book so difficult to follow. Although some of these are minor in themselves, the present reader found their sum total to be more of an obstacle than a mere distraction.

I. Obscurity of the writing.

a) Often whole sections of discussion contain nothing more than string after string of undefended generalization, displaying no clear relation between one sentence and the next (24; 149ff.).

b) The pervasive use of the passive voice creates confusion and forces the reader to re-read sentences. Take the first two sentences of the introduction as an example of this frequent lack of precision:

There has been considerable growth in the understanding and estimation of Stoic logic in the last thirty years, yet an important dimension of this Stoic achievement has not been grasped. Stoic logic was broadly conceived to include their theories of knowledge and perception… (1).

Given that the first sentence alludes to scholars’ understanding of Stoic logic thirty years ago, it is natural for the reader to assume that the people doing the conceiving in the second sentence are those very scholars. Of course, upon completing this sentence, it becomes clear that Barnouw is now talking about the Stoics themselves. But certainly the author’s point would have been crystal clear from the start had he just avoided the passive voice and specified who he was talking about (“The Stoics conceived their logic to include…”).

c) Metaphorical and idiosyncratic language: Plato’s concept of phantasia“opens the way” for the Stoic view (3); certain issues lie “coiled up” inside Plato’s coinage of phantasia (10); Plato has “saddled” phantasia with an entanglement with judging (10); propositions are “underwritten” by a percipient (16); the continuity of perception with sign and proof is the “axis” of Stoic logic (153).

d) Incoherent sentences: “Perception, for the Stoics, anticipates the picking out and linking of elements and aspects of reality and their assertion as real that are accomplished in the linguistic proposition” (3). “Based on a relation of attribute and subject which has always been found to hold, e.g. a woman’s giving milk taken as a sign that she has given birth, but which could conceivably happen without that preceding condition in the sense that it is something that could be otherwise” (139). “Sextus quotes the last definition cited above, bringing out the contrast which its supporters take it to represent to a kind of formalist approach” (276).

e) Verbosity: “In Adversus Mathematicos Sextus interrupts the presentation just underway with a distinction between a common and a private or particular… sense of ‘sign’, in language which refers to the non-evident… in a way which ignores and diverges from the distinction that is about to be made, since the common meaning which anticipates the recalling sign is not defined by reference to the temporarily non-evident, while the particular sign is endeiktikon of the non-evident without qualification” (224). “In the summation of this dichotomously proceeding definition…” (310).

f) Frequent use of verbal nouns: “[The continuity between phantasia and doxazein ] is of interest here for its underscoring of the problem of perception” (19). “…the interdependence of Sceptic and Stoic can be observed… in their working out the characterizations of perception…” (170).

g) Redundancy: “Recalling sign and revealing sign are taken to be opposed even by the contending sects in the debates of the schools of Hellenistic medicine….” And then, three paragraphs later, “The contrasting types of sign were pitted against each other even in the debates of the medical sects…” (216); “declarative or assertoric predication” (13); “technical logical formulation” (234); “a broader, laxer conception” (307).

II. Inconsistent and inaccurate editing:

a) The title of chapter four is inconsistent with the header for that chapter.

b) Inconsistent and incorrect transliterations of Greek: both ” phantasiai” and ” doxas” as plural (20); “somethings” as tini (34, 156); the oblique cases used in transliteration, e.g., ” to endiathetô logô” (without iota subscript!), instead of endiathetos logos (154-55); o for ô, e.g., ” ex heauton / heteron” (168), and ” huparchonton” (192).

c) Typos: “qurestions” (126); “Aristotle does develops” (134); “yet what it reveal appears” (170); “It is import to recognize” (277); “a complete argument is constructed out a conditional” (277); confusion with commas and periods (98, 275, 313, 383); erratic use of italics (51, 97, 342); superfluous spaces between words throughout the book.


1. In his concern for historical development and transition (cf. 151-52), Barnouw’s approach differs from another recent book on Aristotelian and Hellenistic theories of the sign, whose author emphasizes that the relevant theories and debates “cannot be made to fit the pattern of a single continuous development in which positions are taken and defended with reference to a framework common to all parties” (Allen, J., Inference from Signs: Ancient Debates about the Nature of Evidence. Oxford: Clarendon Press, 2001, pp. 7-8).

2. Or more precisely, Barnouw is concerned with Chrysippean logic, as he explains on pp. 152-53.

3. See p. 153 for the clearest statement of this thesis.

4. “The Origins of Non-deductive Inference”, in J. Barnes, et al. (eds.), Science and Speculation: Studies in Hellenistic Theory and Practice. Cambridge: Cambridge University Press, 1982, pp. 193-238.