Logoi and Pathêmata

| 13 →

I. Introduction

1. Logoi and pathêmata

This passage at the beginning of Aristotle’s DE INTERPRETATIONE has been labeled as “the most influential text in the history of semantics.”¹ It seems to be the first time in the history of philosophy that someone has assumed a necessary third level of representations in-between language and reality which is language-independent. This statement might be read as demanding a theory of concepts as mental representations.² The proposed representations are said to be a subject matter of psychology as understood by Aristotle. They are closely tied to his theory of perception and imagination. Pathêma in this context might be equated with the notion of phantasma in the psychological writings.³ However, pathêmata/phantasmata seem to have a role in Aristotelian logic as well since Aristotle mentions them in this particular context before he starts to explain how complex linguistic representations can be true and false.

Pathêmata have a different relation both to simple words (onomata) and the external things (pragmata) they represent. The words are symbols of the pathêmata. Their relation is accidental, mediated by convention. In contrast, pathêmata are signs of the external things by virtue of being their likenesses (homoiômata). Thus, the beginning of DE INTERPRETATIONE is the natural starting point for research on Aristotle’s notion of concept if he has such a notion. As a ← 13 | 14 → next step, the psychological writings and the notion of pathêma (or phantasma) developed therein must be consulted. One central aim of this book is to reconstruct what Aristotle had to say about the kinds of entities that later became known under the labels “concept,” “conceptus,” “Begriff,” “idea,” etc. Hence, the theory presented in this few lines and its links to the psychological writings will be one important source and point of reference used for this purpose.

However, there are some other contenders for “concept” in Aristotle. Horos from the logical writings has often been translated as “concept”4, as well as other expressions such as noêma.⁵ In this dissertation, the focus will be on yet another expression which has often been translated as “Begriff” or “concept”: logos.⁶ This usage of logos (which is a highly ambiguous word) can be found foremost in the logical writings but also throughout the whole Corpus Aristotelicum. Logos is favored over the other possible candidates in this context since there are passages where it seems to have exactly the same function as pathêma, i.e., designating a representational level between language and reality. Nevertheless, the other contenders will also receive due attention in the later parts of this dissertation. Their relations to pathêma and logos will be clarified.

The most prominent passage of this kind, in which logos is used to designate an intermediate level of representation comparable to a conceptual level, is located at the beginning of CATEGORIES:

Words (onomata) are said to have a corresponding logos, or more precisely a logos tês ousias. In the case of homonymy, the word is the same while the logos is ← 14 | 15 → different. If one wants to adopt the parlance of concepts, one could say that the word is connected to two different concepts. Consequently, it has two different meanings. Logoi seem to play the same role here that pathêmata do in DE INTERPRETATIONE. They are independent of words and can remain the same while their corresponding words change. The Greek word logos is notoriously slippery and it should be noted that I am referring to logos as it is used in this passage of the CATEGORIES throughout unless stated otherwise. Logos in this passage has often been understood as ‘meaning.’ However, the notion of meaning is, in itself, not without ambiguity. In some uses, it seems to be equivalent to some meanings of the modern term “concept,” in others it is not, e.g., meaning as extension. Logos may rather be translated as “definition”8 in this and similar contexts. On the other hand, Aristotle has a separate technical term for definitions, i.e., horismos.

In summary, neither pathêma nor logos can be easily defined taking solely these two passages into account, but both seem to be central notions to the Aristotelian theory of language, thought, and concepts. Because they are not clear notions in isolation, their relation to each other is even less clear. Thus, to answer whether Aristotle has a theory of concepts and what it amounts to it is necessary to answer the following three questions first: a) What are logoi? b) What are pathêmata? c) How are both related to each other?

2. Theories of concepts

It is very difficult and presumptuous to speak of theories of concepts in general. This cannot presuppose that these theories are homogenous in any way. The meaning of the word “concept” is highly ambiguous both in everyday use and scientific language.⁹ If one wants to speak of them as homogenous, the easiest approach would be to focus on the technical term “concept” and its relatives such as Latin “conceptus” and maybe also German “Begriff” which is a direct terminological translation of “conceptus.”¹⁰ ‘Concept’ is a central notion in modern ← 15 | 16 → philosophy.11 For this reason, it is surprising that there is no equivalent term for “concept” in Plato or Aristotle.¹² It is not restricted to philosophy but can also be found in psychology, linguistics, and other disciplines. The usage of “concept” in the other disciplines builds to some degree on philosophical theories of concepts. In addition, more recently psychological theories of concepts have been the subject of discussion in the modern philosophy of mind. How “concept” is defined in psychology and linguistics depends mostly on demands imposed by the specific discipline. Nevertheless, it can be interesting to compare theories from linguistics and psychology to older philosophical theories of concepts.¹³ After all, there were no separate disciplines for these subject matters throughout most of the history of philosophy. Brewer (1993) states similar reasons for why it is important to consider philosophical theories of concepts in his typological account of psychological theories:

The philosophical problem of concepts is closely connected to the problem of universals.14 Plato can be considered the first philosopher to note that some things are said universally, e.g., generic terms or property terms. However, even though there is no corresponding entity in the world, these terms do make reference to something. We actually do speak as if general terms and property terms had real referents, e.g., as though justice was a real entity. Plato’s solution was to assume that universals actually exist as independent entities (ideai or eidê).

There are three classical solutions to the problem of universals posed by Plato. These solutions have been canonized in medieval philosophy as universals existing either ante rem, in re, or post rem. Universals might be mind-independent higher-level entities like Platonic ideas. They might be present in particular things and real because they are the inherent forms of particulars. This solution to the problem of universals is sometimes described as Aristotelian. Finally, universals might only exist in human minds or in human language. The former of those two positions has been called conceptualism (i.e., universals are mental concepts) while the other has been called nominalism (i.e., universals only exist as general terms in language).¹⁵

Mental representations of universals have been called “concepts.” This is the most relevant usage of the term in the context of this dissertation. If pathêmata qualify as concepts, Aristotle can be described to have a theory of concepts as mental particulars. Additionally, it should be noted that the term “concept” has been used in different ways. For example, it stands for Platonic ideas, linguistic meanings, logical primitives, and many more. In contrast, modern philosophy is mainly concerned with questions about the ontological status and the epistemic role of concepts. For example, the question as to how concepts can be shared, if ← 17 | 18 → they are conceived as mental particulars, as well as the question of how concepts can be “about” the world and, therefore, correctly represent it.

Theories of concepts in modern psychology treat concepts as mental representations of real kinds16 (sometimes including mental representations of particulars, i.e., individual concepts). Psychology is not primarily interested in ontological or epistemological questions but rather in how concepts are able to fulfill specific cognitive tasks; e.g., how is one able to categorize something as belonging to one kind and not another? What information is included in conceptual representations? How are they structured? And how can higher-order cognitive operations (e.g., deduction, induction, plan-making) be described?

Linguistics is interested in the study of language on all levels. Some of these levels are immanent to language as an empirical phenomenon, e.g., phonetics, morphology, and syntax. Semantics, on the other hand, discusses the phenomenon of linguistic meanings. Word meanings can be equated with concepts, i.e., if something qualifies as a word meaning, it also qualifies as a concept.¹⁷ Consequently, semantic research starts with observations on language use. Alternatively, accounts in linguistics can make reference to psychological theories of concepts, i.e., they propose that linguistic representations do have a meaning by virtue of their connection to language-independent mental representations which are used in categorization tasks and other higher-order mental operations and are the focus of research in psychology of concepts. This is roughly equivalent to what Aristotle seems to be saying in DE INTERPRETATIONE.

Aristotle is an interesting candidate to compare to modern theories of concepts since he is the first philosopher with an extensive corpus and an elaborate and mostly coherent philosophical system. He is also the first that clearly states the necessity of a mental representational level in between the level of language ← 18 | 19 → and reality. Furthermore, he is said to be the predecessor both of classical theory (one of the earliest psychological theories of concepts) and of modal theories of concepts (one of the most recent developments in the psychology of concepts) at the same time. If Aristotle is to be compared to modern theories of concepts, as I propose, it is necessary to first answer the following questions: a) To which theory of concepts is Aristotle compared? b) Can Aristotle be said to have a theory of concepts? c) Which notion/technical term in Aristotle comes closest to “concept” in the modern debate? Possible answers for the last question have been mentioned above and will be discussed in more detail below.

3. Aristotle and the modal/amodal distinction

In the last decade, there has been a trend to see cognition not as symbolic manipulations in analogy to computer programs (which was the paradigm for early cognitive science) but rather as an embodied process which takes place in a biological body and makes use of the specific features of the cognitive agent’s body. This trend also includes embodiment theories of conceptual representations. One of the most influential proponents of concepts as perceptual representations, i.e., representations inherently connected to the perceptual brain systems, is the cognitive psychologist Lawrence Barsalou. His theory is inspired by experimental results measuring the brain’s activity during conceptual tasks. He proposes a systematic and general theory of how a purely perceptual system of cognitive representations that is strong enough to account for the variety of conceptual tasks can be conceived. One of these requirements is that conceptual representations resemble actual and concrete perceptions. Barsalou’s account has received a lot of attention in contemporary philosophy –both negative and positive.¹⁸

Concerning his predecessors in philosophy, Barsalou claims that actually most thinkers throughout the history of philosophy have been proponents of a perceptual theory of higher cognition, in contrast to commonly held beliefs. He cites Aristotle as one of the earliest proponents:

This historical remark seems to disregard the strong rationalist tradition in philosophy originating with Plato and continued by thinkers such as Descartes and Frege in later periods. Generally speaking, rationalists denied the epistemical value of perception. Barsalou’s historical evaluation is at least an oversimplification. However, it still cannot be ruled out without consideration that Barsalou might be right, e.g., about Kantian schemata and especially about the English Empiricists – in fact his own position has often been called “Neo-Empiricism” by philosophers.

It might seem anachronistic to compare Aristotle to very recent theories of concepts, especially psychological theories. On the other hand, there was renewed systematic interest in Aristotle and his philosophy in the last century after he had been mostly rejected by philosophers in the centuries before. Höffe (1996) still sees a lot of potential for innovation in Aristotle’s works. He cites Brentano, Heidegger, Łukasiewicz, and, most prominently, modern virtue ethics (such as Anscombe) as being directly influenced by Aristotle. In addition, he suggests that there are strong parallels between Aristotle’s philosophical method and modern analytical philosophy. Furthermore, he connects Aristotelian biology to the epigenesis theory in modern biology, which states that phylogenetic development is mirrored by ontogenetic development.19

The Aristotelian notion of pathêma does have a lot in common with Barsalou’s account of perceptual symbols: conceptual representations are said to build on more primitive representations which originate in perception. Furthermore, pathêmata are said to bear likeness to external things. Barsalou, of course, would not agree to this. Pathêmata are likenesses of the pragmata by virtue of being likenesses of perceptions which, in the Aristotelian framework, can meaningfully be described as likenesses of the pragmata. It is only this second relation that Barsalou (and most modern psychologists) would deny. The similarity relation between perceptions and conceptual representations is a central and constitutive element both in Barsalou and in Aristotle. All that has been said up to this point might suggest that it is reasonable to accept Barsalou’s evaluation and attribute a perceptual theory of concepts to Aristotle. ← 20 | 21 →

Barsalou’s perceptual symbols are conceived as modal, i.e., tied to modality-specific systems of sense perception.20 This contrasts with amodal theories of concepts which assume concepts to have an independent representational format somewhat like a computer code. Those arbitrary symbols are embedded in language-like structures, e.g., in schemata or frames, which can be manipulated by a cognitive system. Thus, amodal theories of concepts are sometimes referred to as “propositional” theories of concepts.²¹ Those theories are presented as a clear-cut alternative to perceptual/modal theories of concepts. The most radical example of this kind of theories is a classical or feature list approach to concepts in which concepts are understood as an unstructured list of features, while the features themselves are arbitrary symbols. In addition, concepts are understood as definitional, i.e., the features contained in a concept are all necessary and jointly sufficient for something to be subsumed under a given concept. The two aspects do not always have to be connected; in other words, probabilistic and non-definitional theories of concepts might be symbolic and amodal. At first sight, this description of concepts seems to fit what Aristotle discusses under the labels of logoi and horismoi in the logical writings. Several authors see these similarities and mention Aristotle’s theory of definitions as the first example of a classical and amodal theory of concepts.²² Brewer (1993) even uses the term “Aristotelian concepts” to designate classical theories of concepts which are based on the belief that members of a category share necessary essential features.²³

If both accounts are accepted, Aristotle seems to have both a modal and an amodal theory of concepts. This should be an inconsistency according to proponents of modal theories of concepts. Furthermore, in addition to the internal questions already imposed by Aristotelian philosophy, this comparative perspective brings up new questions: a) Can modal and amodal theories indeed both be linked to the Aristotelian notions of pathêma and logos respectively? If Aristotle has both a modal and an amodal theory of concepts in his philosophy: b) How are both related to each other? c) Is the distinction between modal and amodal representations perhaps not as clear-cut as proponents of modal theories seem to indicate? d) Can the relation between logoi and pathêmata in Aristotle be transferred ← 21 | 22 → to modern theories, i.e., is it possible, in a modern context, to be a proponent of a modal and an amodal theory of concepts at the same time in the same way Aristotle was? e) In addition and more generally, it can be asked, whether Barsalou’s historical evaluation is correct in regard to other philosophers.