Concepts before Kant: Objects of thought through the history of philosophy
Name: Jos van Leeuwen Student number: 10190899
Master’s programme: Philosophy of the Social Sciences Supervisor: Hein van den Berg
Second reader: Jacques Bosch Publication date: 24-9-2018 Word count: 32678 words
1
Index
Abstract ………... 2
Introduction ………..………. 2
Method ……. ………..………. 7
1 Concepts in classical philosophy ....………... 14
1.1 Plato .. ……….………..……… 14
1.2 Aristotle .…….………..……… 20
2 Concepts in Hellenistic philosophy ..……….. 27
2.1 Epicurus …….………..………. 27
2.2 Stoicism …….………..……… . 33
3 Concepts in early Enlightenment philosophy ……….. 40
3.1. The origin of the term idea ………... 40
3.1 Descartes …….………..……… 42
3.2 Locke …….………..………. 49
4 Concepts in late Enlightenment philosophy………. 55
4.1 Hume …….………..……….. 55
4.2 German Rationalism …….………..….. 61
5 Concepts according to Kant ………..……. 68
Conclusion …….………..……….... 76
2
Abstract
In modern philosophy, ‘concepts’ are conceived of as general, abstract mental
representations. This view is often traced back to Kant. However, similar entities were already recognized by ancient, medieval and early modern philosophers. These include the ‘simple thoughts’ of Aristotle and the ‘ideas’ of Descartes and Locke. In this study, I investigate developments in the theory of concepts from Plato to Kant using the methodology of the history of ideas. More specifically, I construct models of the views of concepts put forward by different philosophers, and then assess the similarities and differences between their views by comparing their respective models. I argue that the most important properties of concepts were recognized by every philosopher which has been investigated in this study, including their status as mental objects, their representational nature, and their ability to represent multiple individuals at the same time. This means that Kant’s contributions to the theory of concepts are less substantial than it has been argued.
Introduction Concepts and sensations
References to ‘concepts’ are a staple in modern philosophy and science. In logic, they are the elements of propositions. In philosophy of science, they are the building blocks of scientific theories. In epistemology, they are representations of general properties. In linguistics, they give words their meaning. And in psychology, they are what allow us to construct mental schemes, and thus make sense of the world around us. This ubiquity of the language of concepts is nothing strange, for the concept of a ‘concept’ fits very well with some common distinctions made in the aforementioned fields. For instance, the distinction between concepts and words allows distinguishing somethings meaning from the way in which it is expressed, and the distinction between concepts and objects allows distinguishing thought and being, representation and reality, epistemology and ontology. In this thesis, I am primarily concerned with a third distinction and the view of concepts it entails, namely that between concepts and percepts, sensations, or impressions. This means that this thesis is about ‘psychological concepts’ employed in thought, rather than ‘logical concepts’ existing in some abstract realm.
Like the concept-word and concept-object distinctions, the concept-percept distinction is related to other philosophical oppositions, such as between thought and sensation, between mediate and immediate representation, between discursive and pictorial knowledge, and between intelligible and sensible properties. And like the other distinctions, the
concept-3
percept distinction seems self-evident to the point of triviality. For does not everyone realize that most of our thought goes beyond what we directly perceive? We see a small yellow blob in the sky, but we think about the sun which is hundreds of times bigger than the earth. We see a blue fluid substance, but we think about water and how it is able to both freeze and evaporate. We see a black square here and a wood-colored square there, but we know that the one is a couch and the other a table, that one is for sitting on and the other for putting things on, and that both are furniture. Therefore, all it takes to make a distinction between concepts and percepts is to notice the difference between what we see and what we think about what we see, distinguish these two activities from each other as ‘sensation’ and ‘thought’, and then divide the elements of these activities into ‘percepts’ and ‘concepts’ accordingly.
The basic nature of these distinctions is shown by their long and venerate history. Indeed, from the time of the Ancient Greeks on, philosophers have contrasted thinking and perceiving, understanding and sensibility, intelligibles and sensibles, the evident and the hidden. However, one central aspect of this opposition is conspicuously absent during much of the history of philosophy, or at least not explicitly entertained. This is the distinction between the mental objects or elements involved in all of these activities—that is, between concepts and percepts themselves. Though the distinction between thought and sensation comes close, it refers to complex mental processes and not to the elementary representations these involve. Similarly, the distinction between sensibility and understanding posits two different mental faculties, but does not necessitate that these involve different kinds of mental objects. But if the distinction between concepts and percepts is not as old and basic as its fundamental status in modern philosophy seems to suggest, then what is its source? Why was it made? And why not before?
Kant's distinction between concepts and intuitions.
One of the most famous versions of the distinction between concepts and percepts was provided by Kant. In his famous Critique of Pure Reason (CPR), he introduces this distinction in the following way: “objects are given to us by means of sensibility, and it alone affords us intuitions [Anschauungen]; but they are thought through the understanding, and from it arise concepts [Begriffe]” (CPR. B33). But besides having a different source, intuitions and concepts also relate to the objects they represent in fundamentally different ways: “no representation pertains to the object immediately except intuition alone. Thus, a concept is never immediately related to an object, but is always related to some other representation of it, whether that be an intuition or itself already a concept”. The first are images of the things
4
around us, which rest on ‘affections’, the second are rather rules for ordering such
representations, or ‘functions’ (ibid., B93). In other words, while intuitions are immediate and perceptual, concepts are mediate and discursive.
Kant’s distinction between concept and intuition has been widely influential. Thus, Peirce wrote that the “greatest merit of Kant's doctrine lay in his sharp discrimination of the intuitive and the discursive processes of the mind” (KRI.; 15), while Tolley notes that “Kant's distinction between intuitions and concepts has shaped, and continues to shape, much of the discussion in contemporary philosophy of perception” (Tolley 2013: 108). Bolzano credits Kant with popularizing this distinction as well, arguing that “although Kant was not the first to have drawn attention to the distinction between intuitions and concepts, he caused it to be generally recognized” (TS.; 104).
Others even claim that Kant was the first to give the term ‘concept’ itself, or its German equivalent Begriff, its modern meaning. Thus, in the Dictionary of Untranslatables it is claimed that “it was with Immanuel Kant that Begriff acquired a specific philosophical meaning” (Buttgen, Crepon & Laugier 2014: 90), and in his Hegel Dictionary, Inwood states that the meaning of Begriff “was stabilized by Kant” (Inwood 1992: 58). Similarly, Murray argues that “much of our current understanding about concepts has come from Enlightenment philosophers, particularly Kant” (Murray 2017: 85), and according to Roth, Kant “described and theorized the concept of concept in the way that it can be found in philosophy and psychology to the present day” (Roth 2011: 109).
Such assessments seem to imply that both the idea of a ‘concept’ and its distinction from other mental states first arose with Kant's philosophy. And indeed, a short survey of his predecessors seems to prove this. Plato and Aristotle distinguished sensation and thought, but never characterized this distinction in terms of their elements. Descartes and Locke
indiscriminately spoke about ‘ideas’, subsuming every mental object under this very broad and general term. Hume at least distinguished ‘impressions’ and ‘ideas’, but regarded both as mental images, while the German Rationalists used terms like conceptus and Begriff without distinguishing between mental objects involved in thought and sensation.
However, in contrast to some of his other famous distinctions, like those between apriori and aposteriori cognition, between analytic and synthetic judgements, or between appearances and things-in-themselves, Kant does not provide a proof or even a rationale for opposing concepts to intuitions in this manner. Neither does he indicate how his terminology corresponds to or departs from that of earlier philosophers. This was already recognized by Kant's contemporaries and immediate followers. Thus, Schopenhauer criticized Kant for “not
5
saying a word about what thought is, what the concept is, what is the relation of abstract and discursive to concrete and intuitive, what is the difference between the knowledge of men and that of brutes, and what is reason” (WWR.; 84). Similarly, Reinhold argued that because Kant had not determined the concept of a “representation in general”, his distinction between “sensible representation, concept, and idea” remained ungrounded (FPK.; 82).
The source of Kant's distinction
It appears that although the source of the modern distinction between concepts and percepts (or intuitions) is to be found in Kant, his account of this distinction is less than informative about why he made this contrast in this way and why he thought it to be an improvement over earlier accounts. However, it is well known that Kant's theory of cognition was heavily influenced by the metaphysical and logical works of his immediate predecessors, like Baumgarten and Meier, who in their turn were followers of Leibniz and Wolff. And indeed, all these philosophers provided elaborate (though often mutually inconsistent) taxonomies of representation, contrasting ‘notions’, ‘ideas’, and ‘concepts’ on the basis of their generality, immediacy, distinctness and so on.
Kant’s immediate predecessors, in their turn, realized that they were not the first who were engaged in making such distinctions. Thus, Baumgarten wrote that “the Greek
philosophers and the Church fathers have already carefully distinguished between things perceived and things known” (RP.; 78). Another important point the German philosophers already grasped is that such distinctions transcend all philosophical schools. For instance, Reinhold claimed that concerning the distinction between thinking and sensing, “Plato was just as much in agreement with Epicurus, and Zeno with Aristotle, as Leibniz later was with Locke, and Mendelssohn with Helvetius” (FPK: 255). Modern scholars have traced back the theory of ‘concepts’ beyond Kant as well, including not only Enlightenment thinkers such as Descartes and Locke (cf. Machery 2009), but also ancient philosophers like Plato, Aristotle (cf. Helmig 2012), Epicurus, and the Stoics (cf. Knuuttila & Sihvola 2013).
Still, no general overview does yet exist of the development of the theory of concepts through the whole history of philosophy, and neither has the relation between Kant’s view of concepts and earlier accounts received much attention. Weitz’ (1988) Theories of Concepts comes close, but focuses on logical rather than psychological concepts. Furthermore, it does not include philosophers from late antiquity, even though these had a strong influence on Renaissance and Enlightenment thinkers. Knuuttila and Sihvola (2013) discuss the development of the theory of concepts as well in their Sourcebook for the History of the
6
Philosophy of Mind. However, these authors stop just short of Kant, which means that it again
remains unclear how Kant’s theory of concepts is related to earlier accounts.
Research question
In summary, it seems that Kant provides neither a history nor a justification for the way in which he distinguishes concepts from other mental objects. However, the importance of Kant’s view of concepts both for modern theories of cognition and for Kant's critical philosophy cannot be denied. Also, it is clear that Kant did not operate in a philosophical vacuum, but was heavily influenced by his predecessors. Still, the origin and sources of Kant’s theory of concepts are only summarily treated in earlier literature, which creates the image that Kant was the first to refer to concepts in a modern way.
In what follows, I investigate the extent to which Kant's view of concepts, including his distinction between concepts and other kinds of mental objects such as percepts,
sensations, or intuitions, can be found in earlier philosophers, either explicitly or implicitly. In this way, I want to ascertain the origin and development of the concept of a concept itself. I approach this subject primarily from a descriptive angle, investigating the writings of important philosophers of cognition through history only to reconstruct their theories of concepts and note their similarities and differences. However, I also assess the extent to which their theories can take account of important problems for the philosophy of cognition, such as the distinction between discursive and pictorial cognition, the nature of concept formation, the possibility of representing general kinds, and the ontological commitments their solutions entail. Such an investigation belongs to a branch of philosophy called the ‘history of ideas’. In the next section, I explain more into depth what kind of method this involves.
7
Method History of ideas
How to investigate the development of thought about a certain philosophical topic through history? This is the subject of the discipline called the ‘history of ideas’. According to Lovejoy (1936), its founder, the history of ideas is mainly concerned with ‘unit-ideas’, which he conceived of as the elemental building-blocks of which larger theories and doctrines are composed. Modern accounts of the history of ideas often do not employ the term unit-idea but rather refer to ‘concepts’ (cf. Skinner 1999, Kuukkanen 2008). However, as this thesis is concerned with the very concept of a ‘concept’, I will refer to unit-ideas when discussing the methodology of the history of ideas in order to avoid ambiguity.
According to Lovejoy, unit-ideas take the form of a “a single specific proposition or ‘principle’, together with some further propositions which are, or have been supposed to be, its corollaries” (Lovejoy 1936: 15). After their emergence, the same unit-ideas keep being used by different philosophers while their meaning remains stable. However, they enter into very different systems of thought, gaining new functions and expressions through appearing in new combinations. Lovejoy (1944) argues that the emergence and recession of unit-ideas can be traced through the history of philosophy by isolating them from the substantial theories of which they are a part and then comparing their internal structure. The same idea might then be said to be present in different thinkers if it takes on the same role and has the same
characteristics (Lovejoy 1944: 204-205). This makes it possible to ascertain the presence of the same unit-ideas even in thinkers belonging to very different periods, disciplines and schools. Indeed, Lovejoy himself was convinced that “there is a great deal more that is common to more than one of these provinces than is usually recognized, that the same idea often appears, sometimes considerably disguised, in the most diverse regions of the
intellectual world” (Lovejoy 1936: 14-15).
It might seem rather simplistic to suppose that ideas can be divorced so easily from their context, and then compared and related to each other as if they exist in some theoretical vacuum. However, Lovejoy did in fact recognize that the practice of history of ideas involves a lot more than just constructing formal definitions of basic terms and then pointing to
passages in various authors in which these terms seem to be used in the same way. First of all, the presence of a unit-idea cannot just be read into a certain author. Rather, this is a judgement by the historian about what an author was trying to say, something which often involves translating the specific terminology employed by the author to the general form of the unit-idea. Thus, Lovejoy argued that an adequate interpretation of a text was “dependent upon a
8
knowledge—or an assumption—about what he [the author] was trying to do, which can by no means always be safely or fully inferred from the obvious content of the work” (1940: 13).
Secondly, though unit-ideas in their ideal form are discrete, atomic and atemporal entities, they almost never appear in their pure form. Rather, they are mediated by other forces present in every theory and every field, such as ‘implicit assumptions’, ‘dialectical motives’, ‘metaphysical pathos’, and ‘philosophical semantics’ (Lovejoy 1936: 7-14). For the present project, it suffices to say that the history of ideas must involve a comparison of particular instantiations of unit-ideas with their general form, allowing for the possibility that they deviate from this ideal to a greater or lesser extent. Indeed, most unit-ideas are not as elementary as the term might suggest: they are not simple definitions or propositions, but occur together with others which have become “agglutinated to it” (ibid.: 14). Thus, it has been argued makes more sense to conceive of unit-ideas as ‘models’ consisting of various different parts or ‘components’ rather than just singular statements (cf. Lovejoy 1944: 206, Betti & van den Berg 2014). Lovejoy even argued that in a single author contradictory
elements may be present, for it “cannot be assumed a priori to be impossible for a writer to be attracted by mutually incompatible ideas, to be responsive to conflicting strains in his
intellectual heritage, and therefore to give expression sometimes to one, sometimes to another” (Lovejoy 1947: 421).
Finally, it is not necessary to assume that the presence of a unit-idea in two or more authors is evidence for a direct relation between them. Indeed, Lovejoy himself already distinguished between ‘recurrent’ and ‘persistent’ unit-ideas (Lovejoy 1944: 204). Unit-ideas might emerge and disappear, after which they are rediscovered again after some time.
Furthermore, different authors might develop the same ideas by following the same lines of argument, without ever knowing that somebody else came to the same conclusions.
Issues and solutions
However, even such a more nuanced view of the history of ideas has to take heed of some further issues. One important problem noted by Skinner is that historical agents drop from view. Ideas seem to become independent objects, which ‘emerge’, ‘spread’, ‘linger’, ‘disappear’ and so on (Skinner 1969: 10). This ‘reification of doctrines’ has various
consequences. First of all, because of such reification, earlier thinkers are interpreted from the contemporary point of view, and if they hold comparable views these are regarded as
‘anticipations’ in which the modern doctrine was already immanently present (ibid.: 11). Secondly, it leads to endless bickering about whether a view was already present, or whether
9
it was not ‘fully articulated’ or ‘completely developed’ yet (ibid.: 12). Thirdly, it makes it impossible to understand the significance of a particular idea or doctrine for an author, ignoring the fact that these are used “with varying and incompatible intentions” (Skinner 2002: 84). Finally, it cannot show the importance of a particular idea for an author, but only that it “makes some appearance in his works” (ibid.: 85). Thus, it remains unclear whether the author considered the idea as trivial or important, central or peripheral to their philosophy. In this sense, the history of ideas runs the risk of attributing arbitrary meanings to historical authors, perceiving their writings through some pre-existing ‘mental set’ rather than pointing out real connections between their ideas (ibid.: 58).
Skinner instead argues that the history of ideas must always take the intention of the authors themselves into account (Skinner 1969: 48). This also means that one should not take an author to be concerned with some ‘perennial question’ or ‘Great Conversation’ across history: statements by authors are always “the embodiment of a particular intention, on a particular occasion, addressed to the solution of a particular problem, and thus specific to its situation” (ibid.: 50). This means that a correct assessment of the significance of an author for the development of history depends on the particular context in which they use an idea. To do justice to the particular intentions of the authors I have included in this study, I do not only investigate their references to concepts, but also place these in the context of their more general theories of the mind and of the world.
A second problem, on which Kuukkanen (2008) focusses, is that ideas are historically variable: their meaning continuously evolves, and therefore authors may mean different things when using the same terms. This means that it cannot be assumed that there are stable,
unchanging ideas which keep appearing in the same form. However, this is precisely what Lovejoy’s history of ideas presupposes (Kuukkanen 2008: 354). This also implies a second problem: how to decide whether two authors were even discussing the same idea? Did Plato and Rawls both contribute to the same notion of justice? Did Locke and Rousseau both refer to the same idea of liberty? If definitions become too restrictive, almost no thinkers can be said to have used the same ideas. But if they are too broad, many important differences may be overlooked (ibid.: 355). Therefore, it is necessary to determine in some way the scope of an idea.
Kuukkanen's solution is to take ideas not as discrete entities but as “conventional postulations made by historians for the purpose of describing history” (Kuukkanen 2008: 367). According to Kuukkanen, this solves both the problem of conceptual change and the problem of determining the scope of an idea. Specifically, one should first delineate the
10
aspects of an idea which are shared among thinkers, even if their specific formulation differs. Kuukkanen calls the former the “core” of the idea, and the latter its “margin” (ibid.). This allows highlighting the similarities between the views of different thinkers, while also recognizing the differences which exist. The scope of the idea, then, becomes something determined by the historian, rather than by the authors he investigates. Furthermore, by laying out the core and the margin of a concept, one can investigate not only conceptual stability, but also conceptual change (if the margin changes) and even conceptual replacement (if the core changes) (ibid.: 369-70).
Application and scope
Method
As has been shown, tracing an idea through the history of philosophy involves more than just looking for similarities and differences in the use of similar terms by different authors. Philosophical ideas have multiple properties and relations to other ideas, and are more like models than like single definitions. Their definitions are often left implicit, in which case their internal structure must be reconstructed from their use. They are not independent entities, but are always employed in a specific way by a specific author with a specific intention. They form a part of more general theories and doctrines held by these authors, which can greatly differ even if they use the same concepts. Indeed, authors often give a different meaning even to the idea itself, leaving out certain properties and stressing the importance of others. This means that the historian of ideas cannot presuppose that their preconceived assumptions of what is essential to the idea are correct.
However, this does not mean that the history of the concept of a ‘concept’ cannot be written—just that it is a bit more complex than it might seem. In the following chapters, I investigate whether concepts were already recognized as specific entities by philosophers before Kant, and to what extent Kant’s view of concepts was influenced by such earlier references. On the basis of this investigation, I also develop a model of the concept of a concept. Taking heed of the precautions mentioned above, I do not view earlier theories as ‘anticipations’ or ‘prefigurations’ of Kant’s view of concepts, as they might have proceeded from different intentions and different conceptualizations. Neither do I look for singular features or explicit definitions of ‘concept’, nor do I determine beforehand which
characteristics belong to its ‘core’ and which to its ‘margin’. Rather, I attempt to ascertain the explicit or implicit role of concepts in the broader theoretical frameworks of specific authors. Thus, I do not presuppose a certain model of concepts, but rather develop such a model
11
throughout this investigation. In this sense, my method is ‘iterative’ and ‘inductive’ rather than ‘deductive’.
Terminology
In the introduction, I have shortly explained what a ‘concept’ is. Here, I give a somewhat more systematic indication of how this term is commonly used to explain what I am looking for in these philosophers. However, I leave a more thorough discussion of the various properties which are ascribed to concepts to the next chapters, where they form the guiding line of the discussion. To preclude the endorsement of any specific conception of concepts, I have looked at various dictionary entries on concepts. Bunnin and Yu’s (2004: 126) Blackwell Dictionary of Philosophy (B), Blackburn’s (1994: 72) Oxford Dictionary of
Philosophy (O), and Audi’s (1995: 169) Cambridge Dictionary of Philosophy (C)—three
important dictionaries of philosophical terms—mention the following characteristics of concepts, set out in the order in which they appear in the respective entries on concepts in these dictionaries:
(M-1) A concept is general: It is “a general notion or idea that may apply to a multiplicity of things and that is expressed by general words” (B), it corresponds to a universal which “can be instantiated by more than one individual thing (or particular) at the same time” (C). Thus, “individuals fall under concepts, and we talk about individuals in terms of concepts” (B) (M-2) A concept is an element of thought: It is “the simplest content of our thinking”, and a “component of propositions” (B).
(M-3) A concept classifies objects: It serves as a “principle of classification, something that can guide us in determining whether an entity belongs in a given class or does not” (C). (M-4) A concept can be more or less general: “Concepts themselves admit a degree of generality. A genus concept is wider in extension than a species concept” (B).
(M-5) A concept is a mental representation: “It mediates between the mind and physical reality, and is a psychological entity with a non-psychological content” (B),
(M-6) A concept resembles what it represents: It serves its “classificatory function by resembling the entities to be classified” (C).
(M-7) A concept does not have a truth-value: “Concepts themselves do not admit of truth or falsity, but the propositions of which concepts are components are the bearers of truth value” (B).
12
(M-8) A concept is used as a predicate in judgments: “A concept is that which is
understood by a term, particularly a predicate” (O). Thus, “to possess a concept is to be able to deploy a term expressing it in making judgements” (O).
(M-9) A concept is required for knowledge: it is “essential to our understanding and knowledge of anything, even at the most basic level of cognition, namely, recognition” (C).
Structure
My investigation of the history of concepts begins with the two greatest philosophers from the ancient world: Plato and Aristotle. Then, I discuss the contributions of Epicurus and the Stoics. After this, I take a leap through history to the early Enlightenment, focusing on Descartes and Locke. Then, I focus on the contributions made in the late Enlightenment by Hume and a group of philosophers called the German rationalists: Leibniz, Wolff,
Baumgarten, Meier, and Knutzen. Finally, I discuss the theory of concepts of Kant himself. Of course, many of these philosophers never referred to ‘concepts’. The word ‘concept’ itself was only coined in the early Middle Ages, meaning that the ancient philosophers which have been included did not even have access to this term. Similarly, Descartes, Locke, and Hume refer to ‘ideas’ rather than ‘concepts’. Only among the German rationalists, the word
‘concept’ and its German translation Begriff are commonly used, though often in a different sense than in modern philosophy. Still, as will be shown, the mental entities discussed by these philosophers resemble the Kantian notion of a concept in many different ways.
Style
In this thesis, I use the following style conventions. Terms which are from a language different than English, such as notio and Begriff, are italicized if they appear by themselves in a running sentence. If they are given next to a normal term in a quotation, they are italicized and put between brackets, except for terms in Greek alphabet, which are only put between brackets. Properties of concepts are identified by a letter and a number between parentheses in boldface, indicating the philosopher who mentions the property and the order in which this property appears in the discussion. For instance, (P-5) stands for the fifth property of concepts mentioned by Plato, and (D-2) for the second property of concepts mentioned by Descartes. Double quotation marks are used when sentences are fully or partially quoted, while single quotation marks are used when single terms or expressions are quoted. I refer to primary sources using standardized abbreviations and page numbers which are common among
13
do not cite a work again if the same page of the same work is cited multiple times in a row. The last citation from a page is indicated by a single (ibid.). I only cite a work again if a reference is from a different page.
14
Chapter 1. Concepts in classical philosophy
The philosophers I investigate in this chapter are Plato and Aristotle. Though most of their specific doctrines and arguments have been superseded, these two giants of ancient philosophy have in many ways remained relevant until the present day. Indeed, their writings on metaphysics, epistemology, logic, ethics and politics have served as a foundation for all subsequent work in these areas. Whitehead even famously claimed that “the safest general characterization of the European philosophical tradition is that it consists of a series of footnotes to Plato” (Whitehead 1929: 39). However, a topic the ancients wrote about which receives less attention in present philosophy is their psychology, including their theory of concepts. To a certain extent, such neglect seems justified, as they seem to lack a general theory of mental objects. However, in this chapter, I argue that Plato and Aristotle did in fact already recognize such entities, and that these resemble modern concepts in various ways.
1. Plato
Traditionally, scholars have focused on two aspects of Plato's philosophy: the theory of eternal Forms and the theory of the immortal soul. These are deemed of such importance that they are often referred to as the “Two Pillars of Platonism” (cf. Cornford 1934: 2). For the most part, Plato's view of the human mind only figured in discussions of the theory of knowledge through recollection, sometimes called the “Third Pillar of Platonism” (Miles 2003: 259). None of these three privileged subjects seems particularly relevant for the study of concepts. However, as Weitz (1988) notices, it is clear that Plato employed many particular concepts, though his theory of concepts is implicit, “to be elicited from the way or ways in which he employs the particular concepts he deals with in his writings” (Weitz 1988: 1-2). In the following, I use this procedure to argue that concepts do in fact play a role in each of Plato’s three central doctrines. First, I argue that Platonic Forms seem to resemble modern concepts in multiple ways, although Plato himself explicitly rejects their identification. After this, I discuss the role of concepts in Plato's theory of recollection, and how he regards this as a proof for the immortality of the soul. Finally, I focus on Plato's theory of the commons, which like the Forms are said to be accessible only through thought.
Platonic Forms
In brief, Plato's theory of Forms argues that beyond the sensible world of everyday experience there is a world of independently existing ideal objects which are called Forms or Ideas. These ideal objects are invisible, incorporeal, eternal and unchangeable, which means
15
that they can only grasped through the mind by itself. Sensible objects, in contrast, are merely deficient copies, imitations, or images of the Forms, and exist only because they participate in ideal objects. Thus, according to Plato only the Forms are truly real. Lastly, one and the same Form, such as Beauty, Justice or Being, might be present in many different sensible objects (cf. Audi 1999: 315, Bunnin & Yu 2004: 264).
The Forms are not only central to Plato's ontology, but also to his theory of cognition. Essentially, he recognizes four types of thought or “affections in the soul [παθήματα ἐν τῇ ψυχῇ]” (Rep. 511b): conjecture, belief, understanding and reason. These are distinguished by the extent to which they are able to grasp the Forms. Conjecture and belief are concerned with visible objects: the former with “images [εἰκόνες], shadows, and reflections [φαντάσματα]”, the latter with the things of which these are images: “the animals about us and all plants and the whole class of objects made by man” (ibid., 510a). Understanding and reason, however, are able to reach beyond the visible world to what is intelligible. The objects of understanding are “the things imitated in the former division”, but now “treated as images [εἰκόσιν]”
themselves, while reason “relies on Forms [εἴδεσι] only” (ibid., 510b).
Forms and concepts
At first sight, Platonic Forms seem a lot like concepts. As Lotze (Log.; 436) argues, both Forms and concepts are not dependent on contingent sensible experiences for their meaning, but are essentially stable entities. Likewise, Schopenhauer (WWR.; 109) notes that both a Form and a concept may without any problem be applied to many different individuals. For instance, the concept ‘human’ applies not only to Socrates, but also to Plato and Aristotle, and the Form of Beauty is instantiated in all beautiful things, whether natural of artificial, living or nonliving. In this sense, both are universals which constitute ‘the one over the many’ (cf. Ross 1955: 24, Matthews & Cohen 1968: 634).
Indeed, in the Parmenides Plato himself raises the possibility that “each of these Forms may be only a noema [νόημα]” (Prm. 132b), henceforth translated as ‘thought’. The reason for this suggestion is that both in the case of thoughts and Forms, “one” is “set over all of them . . . ever being the same over all”. Parmenides, who serves as the interlocutor in this dialogue, offers two arguments against this theory. First, thoughts must always be “of something”, and more specifically about “something which is”. In the language of modern philosophy, they are intentional, directed to something external to themselves. Furthermore, they are objective, directed to something which has independent being. However, if Forms were thoughts, then thoughts about Forms would not have this property: they would be about
16
themselves, and thus “about nothing” (ibid.). Secondly, everything participates in Forms. However, if Forms are thoughts, then “everything is made of thoughts”, and thus either
“everything thinks, or, although thoughts, they are without thought”—both possibilities which Plato rejects (ibid., 132c). Though these arguments may seem not particularly strong—indeed, according to Helmig it is “somewhat exaggerated to call Parmenides’ reply a refutation of conceptualism as such” (Helmig 2012: 51)—they are enough for Plato to reject this theory.
In this short passage, Plato mentions multiple important characteristics of noemata or ‘thoughts’. A thought here is conceived of as a stable entity which is “ever the same over al”, not as a dynamic process or occurrence in the mind. Thus, in this passage a thought seems to refer to a mental object rather than a mental act (P-1). In this sense, it is contrasted with the process of thought, i.e. thinking, and the faculty of thought, i.e. the mind, which are referred to by respectively the terms noesis and nous. However, its connection to these terms does indicate that such a mental object is connected to thought and understanding, rather than to other mental faculties like sensation or imagination (P-2). Such ‘thoughts’ are described as being ‘of something’, which means that they are representations (P-3). This ‘something’ that a thought is about is not just a mental construct or abstraction but “something which is” (P-4). As thoughts are “set over all of them”, they can be applied to many different objects at the same time, which means that they are general (P-5).
Recollection of concepts
The preceding passage is the only place where Plato directly discusses the nature of singular thoughts in themselves. However, this does not mean that concepts are not important for other aspects of Plato's philosophy. Indeed, a closer examination of the theory of
recollection makes clear that it requires the presence of concepts in the mind. In the Phaedo, Plato for the first time argues that our knowledge of the absolute qualities exemplified by the Forms, such as beauty, goodness, justice, and identity, cannot be derived through observation of sensible things, as these never exactly correspond to such qualities (Phd. 74d). For
instance, though we may observe a resemblance between two objects, such as two trees or two humans, these objects are never exactly the same. And though we may observe beautiful objects, these objects always have some imperfection which detracts from their beauty. Thus, we never observe something like complete, absolute identity or beauty.
However, according to Plato we do actually have knowledge of such qualities, for otherwise we could not even refer to something as identical or beautiful (Phd. 74e). Therefore, Plato concludes that there must be another source besides sensation from which this
17
knowledge is derived (ibid.). He proposes that such knowledge is innate to the soul itself, and is ‘recollected’ when we are reminded of it by some experience. What is recollected or grasped, then, is “the ennoia [ἔννοια] which he grasps” (ibid., 73c). Like noema, the term
ennoia is derived from nous. Literally, it means “that which is in one's mind” (Tzamalikos
2013), but it is most often translated as ‘notion’ or ‘thought’ (cf. Helmig 2012, Bonazzi 2017). Its role in the process of recollection makes clear that in this passage a notion is not a mental act but a mental object which exists besides what is actually perceived [αἴσθησιν] (Phd. 73c). For instance, we are sometimes reminded of someone when seeing something which belongs to that person, such as his lyre or cloak (ibid., 73d). Thus, like the ‘thoughts’ mentioned in the
Parmenides, the ‘notions’ mentioned in the Phaedo belong to the understanding rather than to
other mental faculties (P-2). Acquiring such notions involves only reason, without a separate role for sensation (P-6). They are required for knowledge, as knowledge consists in
recollection and what is recollected is thoughts or notions (P-7). Finally, as these notions must be recollected, it is clear that they are located in memory before they are not actively
contemplated (P-8).
Other dialogues in which recollection is discussed are less explicit, and do not mention any specific mental objects or states involved in recollection. And though there are other passages where he refers to noemata and ennoiai, these occur almost always in a
non-technical context. For instance, when asked whether virtue can be taught, Plato gives a quote from the early poet Theognis stating that if “thought [νόημα] could be created and put into a man, many high rewards would they obtain” (Men. 95e). And in a eulogy on Love, Plato says that “song enchants the thought [νόημα] of every god and man (Sym. 197e). Similarly, he once briefly refers to “the notion [ἔννοιαν] of time” (Tim. 47a). However, there are some allusions to the role of mental objects in recollection in other works besides the Phaedo. For instance, in the Meno, Plato states that the results of recollection, the “truths about the things that are”, are “present in our soul” (Men. 86b). And as was already mentioned, Plato describes reason, which is the only mental state which has complete access to the Forms, as an
“affection in the soul” (Rep. 510b).
The commons as concepts
One last dialogue in which mental objects play a substantial role is the Theaetetus, in which Plato attempts to come up with a definition of knowledge. The first option which he investigates is that knowledge is sensation. However, like in the Phaedo, Plato argues that much of our knowledge is not concerned with what is directly present in sensation, but rather
18
with the “common [κοινὸν]”: properties such as being and non-being, identity and difference, and likeness and dissimilarity (Tht. 185c). These are not directly perceived, but rather inferred through comparing and relating different things to each other. Thus, Plato makes a contrast between cognition of the commons, which involves “reflections [ἀναλογίσματα]”, and cognition of more simple sensible qualities such as colors, which only requires mere “affections [παθήματα]” (ibid. 186c).
Some commentators refer to the ‘commons’ and their mental analogues, ‘reflections’, as concepts (cf. Silverman 1990, Frede 1996, Helmig 2012). Others hold that the commons are not concepts but Forms, which like the commons are also present in many particular objects and accessible only to the mind (cf. Cornford 1935: 105, Cherniss 1944: 236n141). However, as Helmig (2012: 25) notes, Forms are always positive entities which really exist, while the commons also include privations like non-identity or non-being. Thus, the former interpretation seems more reasonable.
In any case, it is uncontroversial that these entities can only be accessed by the mind. And though Plato never directly identifies the ‘commons’ with ennoiai or noemata, he does refer to both “perceptions and thoughts [αἰσθήσεσι καὶ ἐννοίαις]” (Tht. 191d) in the
Theaetetus, thus making clear that in this dialogue he again recognizes the existence of
thoughts distinct from sensations. Similarly, when discussing the commons in the Sophist, Plato asserts that they exist “in the soul [ἐν τῇ ψυχῇ]” (Sph. 249c), confirming their status as mental objects. Thus, like the ‘thoughts’ and ‘notions’ mentioned before, and like modern concepts, the ‘commons’ and the ‘reflections’ of them belong to the understanding rather than to other mental faculties (P-2). Further, they represent general properties of objects (P-5).
Conclusion
Though direct discussions of mental objects are rare in Plato's dialogues, they are actually quite important for his philosophy. Indeed, they are the very means through which knowledge of common qualities is accessed through reason, and knowledge of absolute properties (the Forms) can be accessed through recollection. Such ‘thoughts’, ‘notions’ or ‘reflections’ have certain qualities in common. They are mental objects rather than mental acts or powers (P-1). Such mental objects are connected to thought and understanding, rather than to other mental faculties like sensation or imagination (P-2) They are representations of things other than themselves (P-3). The objects they represent exist outside of thought, as the Forms and the commons both exist as independent objects by themselves (P-4). They are general and can be applied to many different individuals (P-5). Their acquisition involves
19
only reason, without a separate role for sensation (P-6). They are required for knowledge, as knowledge consists in recollection and what is recollected is thoughts or notions (P-7). Finally, as these notions must be recollected, it is clear that they are located in the memory when they are not actively contemplated (P-8) These properties are shared with modern concepts, except that modern theories of concepts do not require the real existence of universal objects, and that they do not imply a specific theory of concept acquisition.
However, neither in the passages which have been discussed nor in any other dialogues does Plato offer an extended account of thought and its elements. Indeed, in all his works together, Plato uses the term noema or ‘thought’ only 11 times and the term ennoia or ‘notion’ only 8 times (own investigation). Thus, it remains unclear how a thought which is recollected differs from a sensation which is perceived, or how a thought is originally formed before being recollected in the mind, or what role they play in cognition in general.
20 2. Aristotle
Aristotle’s writings on psychology provided a framework for theories of cognition which, through its influence on medieval Scholasticism, remained unchallenged for nearly 2000 years (cf. Wundt 1912, Kemp & Fletcher 1993, Caston 2012). When discussing the nature of thought, Aristotle focuses mostly on mental operations, their underlying faculties and their external objects. However, sometimes he also mentions elementary mental objects, which when combined yield more complex processes of thought. In what follows, I discuss these thought-elements or ‘simple thoughts’, which correspond to modern concepts in various ways. I first discuss the role of simple thoughts in thought activities. Then, I show how Aristotle compares such simple thoughts to individual words, as both are representations of other things, and both are neither true nor false by themselves. After this, I argue that
Aristotle’s identification of the objects of simple thoughts with essences implies that they are general representations. Finally, I discuss the relation between simple thoughts, sensation, and mental images.
Thoughts and their constituents
Aristotle's most complete account of cognition can be found in On the Soul, which was written around 350 BC. In this work, Aristotle discusses the nature of the soul and its three parts: the nutritive, sensitive and intellective faculties or powers. Aristotle argues that such an investigation must proceed in a certain order: “if one is to say what each power is, then one must rather first say what thinking is and what perceiving is, for . . . operations and activities are prior to their potencies. And prior to these one must have contemplated their objects” (DA. 415a). Thus, the parts of the soul can only be understood after their operations are understood, and these operations can only be understood after the external objects of these operations.
Based on this passage, it might seem that Aristotle is not concerned very much with concepts or other mental objects. However, elsewhere in On the Soul Aristotle does in fact explicitly assert that “mind [νοῦς] is one and continuous, as is thinking [νόησις], which in turn consists of thoughts [νοήματα]” (ibid., 407a), thus including the constituents of thought-activities as a separate level of psychological analysis. As will be shown, in other passages Aristotle makes clear that these constituents, which are often referred to as ‘simple thoughts’, are elementary mental objects which resemble modern concepts (A-1). As their name
indicates, such ‘simple thoughts’ are specifically involved in thought and understanding rather than in other kinds of mental processes (A-2).
21
Thoughts as representations
Aristotle’s most detailed account of the nature of thoughts and their structure is found in On Interpretation, in which he discusses the relation between language, thought, and reality. First, he argues that language involves two levels of representation: “spoken words are symbols [σύμβολα] of affections of the soul [ψυχῇ παθημάτων]; written words are symbols of spoken words”. He then adds that the ‘affections of the soul’ represented by language are also representations themselves: they are “likenesses [ὁμοιώματα] of things [πράγματα]” (DI. 16a).
The word ‘affection’ or pathema which occurs in this passage is the same which Plato used to describe the four types of cognition. At first sight, this term might seem to refer to the passive rather than the active side of the mind, as it is used in other places for sensations (DA. 417b) and other things which happen to subjects (Met. 1061a). However, in the following lines, Aristotle switches to noema, which is often taken to mean that he uses these terms here interchangeably. Indeed, as Sorabji (2012: 631) notes, ancient commentators of Aristotle already interpreted this passage in this way. The word ‘likeness’ (homoioma) which Aristotle uses to describe the nature of such pathemata further implies that these mental states involve a special kind of representation. Unlike spoken and written words, which are merely arbitrary ‘signs’ (symbola), they intrinsically resemble the object they represent. This means that like Plato, Aristotle holds that thoughts are representations of pragmata or ‘things’ (A-3), and that they resemble the things they represent (A-4). In both respects, Aristotelian thoughts are similar to modern concepts as well.
Combined and uncombined thoughts
Aristotle distinguishes representation by words from representation by thoughts by their relation to objects. However, in other respects he regards thought and language as highly similar. After having discussed the representational nature of thoughts and words, Aristotle makes a distinction between uncombined and combined words and thoughts, claiming that “a noun or a verb by itself much resembles a thought [νόημα] which is neither combined nor disjoined” (DI. 16a). This distinction between combined and uncombined thoughts makes clear again that Aristotle recognized the existence of elementary mental objects of which complex cognitions are constituted. Such ‘simple thoughts’ resemble individual words (A-5), and like individual words, they can be combined together to constitute more complex mental acts (A-6).
22
Another important aspect of this distinction consists in the propositional structure of uncombined and combined words and thoughts: sometimes a “thought [νόημα] is
unaccompanied by truth or by falsity, but sometimes it must be one or the other, and so it is in speech; for in composition and division there is truth and falsity” (DI. 16a). Similarly, in On
the Soul, Aristotle argues that if thinking is concerned with “indivisible things [τῶν
ἀδιαιρέτων]” it is never false, but that if it involves a “composition [σύνθεσίς] of thoughts [νοήματα] like a single thing [ὥσπερ ἓν ὄντων]” it is either true or false (DA. 430a, cf. 432a). And in the Metaphysics, he states that “falsity and truth are not in things—the good, for example, being true, and the bad false—but in thinking [διανοίᾳ]; and with regard to simple things [τὰ ἁπλᾶ] and essences [τὰ τί ἐστιν] there is no truth or falsity even in thinking” (Met. 1027b). Thus, like words by themselves, ´simple thoughts´ are neither true nor false in themselves (A-7). In this respect, Aristotelian thoughts again resemble modern concepts.
Simple thoughts and essences
In On Interpretation, the objects of thoughts were described very generally as
pragmata or just ‘things’. However, the passages from the Metaphysics and from On the Soul
specify that uncombined or ‘simple thoughts’ are of ‘indivisible things’, of ‘simple things’, and of ‘essences’, while complex thoughts are composed ‘like a single thing’. To understand what this means, it must be explained what Aristotle means by such terms. In brief, Aristotle argues that ousia or ‘essence’ is the characteristic or set of characteristics belonging to an object which make it what it is. In the Topics, he states that essence “signifies sometimes a substance, sometimes a quality, sometimes some one of the other kinds of predicates”. Thus, attributing an essence takes the form of calling a human individual ‘a man’ or ‘an animal’, a white colour ‘white’ or ‘a colour’, and all other cases in which “either something is asserted of itself, or its genus is asserted of it" (Top. 103b). Such essences are always the same for every object to which this essence belongs. For example, Socrates is a man because he is a rational bipedal animal, and so is Plato. This means it is impossible for something to be added or taken away from the essence, even if, for example, a specific individual has become mad and lost his rationality, or a leg and thus his bipedality. Thus, Aristotle concludes that “the essence is indivisible” (Met. 1034a), making clear that the ‘indivisible things’ mentioned in
On the Soul and the ‘essences’ mentioned in the Metaphysics are the same entities.
Because simple thoughts pick out essences, they always represent real entities. Unlike Plato, Aristotle does not believe in the separate existence of Forms. Rather, they always exist in individual subjects in which they are realized. Each genus thus possesses “certain
23
properties which can be treated as separate even though they do not exist in isolation” (APo. 81b). However, because they occur in real individuals, they are still real entities rather than constructions of the mind. This position is commonly called ‘moderate realism’ (cf. Brakas 1988, Armstrong 1989). And as simple thoughts point out essences, it can be inferred that they represent real entities (A-8).
The fact that simple thoughts are about essences also explains why they cannot be false, for as Aristotle argues, “thought is true about somethings essence [τοῦ τί ἐστι], but not as to every fact about it [τὶ κατά τινος]”. The possibility of falsehood arises only if, in
complex thoughts, essences are predicated of a particular object or of each other—for instance if it is said that “Cleon is white” or that “the white thing is not white” (DA. 430b). Though most modern philosophers do not hold a realist theory of essences, the idea that a concept in itself cannot be false is still common—like Aristotle’s simple thoughts about essences, they do not entail an existence claim.
Simple thoughts as general representations
Like Plato’s Forms, essences are universal entities: they can belong to an unlimited amount of different particulars. Any number of objects can be a man, or an animal, or white, and so on. Though these essences belong to different concrete individuals and are thus realized in a different material substrate, they do not differ in any other way and are thus instances of the same general entity (Cohen 1978: 75). As Aristotle himself notes, “‘man’, and indeed every general predicate, signifies not an individual, but some quality, or quantity or relation, or something of that sort” (SE. 178b).
As uncombined thoughts pick out essences, it seems natural to infer that they are general representations as well. Indeed, in the Metaphysics Aristotle argues that ennoemata, another term used by him for ‘notions’, pick out universals: “art is produced when from many notions of experience [ἐμπειρίας ἐννοημάτων] a single universal judgement is formed with regard to similar things” (Met. 980a-981a). Likewise, in the Posterior Analytics, experience is said to consist of the “universal now stabilized in its entirety within the soul, the one beside the many which is a single identity within them all” (APo. 100a). This makes explicit that experience and the notions of which it consists are more than just collections of particular sensations: they involve universal judgments or representations of objects (A-9). Besides, these passages show that simple thoughts are acquired in an empirical way through sensation (A-10) and stored in memory (A-11), and that after being converted to universal judgments
24
they make knowledge possible (A-12). In all three respects, Aristotle is close to the modern theory of concepts, though the empiricist position is not a necessary part of it.
Simple thoughts and mental images
In the passages discussed above, simple thoughts are described as discursive entities which are much like words. However, unlike words, simple thoughts intrinsically resemble the objects they represent. This points to another view present in Aristotle, according to which simple thoughts are not discursive or symbolic, but pictorial. For instance, in On the Soul Aristotle argues that “the soul never thinks [νοεῖ] without an image [φάντασμα]” (DA.431a). In another passage he even identifies thoughts with images, arguing that “it is by means of the mental images [φαντάσματα] or thoughts [νοήματα] in the soul that one calculates as if
seeing, and that one deliberates on future or present matters” (DA. 431b). And in On Memory, he asserts that “mental images can be contemplated both in themselves and as representations of something else”. Regarding them as representations is like “contemplating a figure in a picture as a portrait” rather than just as a blob of paint on a canvas, but if they are regarded in themselves they “appear to occur as a thought [νόημα], or a mental image” (DM. 450b). This makes clear that thoughts and mental images are closely related, and that the former never occur without the latter (A-13). In this respect Aristotle’s theory of simple thoughts differs from modern theories of concepts, which allow for the possibility of entertaining concepts without visualizing them. However, while images are always of things which can be pictured, thoughts can be of many things, some of which are not sensible, such as the essences
mentioned before (A-14). Thus, there must be some difference between thoughts and images after all.
Aristotle himself asks in On the Soul how the “first thoughts [πρῶτα νοήματα] differ from mental images”. He then offers only a very brief answer: “neither they nor any others are mental images, but they cannot occur without such mental images” (DA. 432a). Thus, though simple thoughts do in fact differ from mental images, they do not occur independently of them. This is not very informative, but in another passage Aristotle describes how the mind employs images: it “thinks the forms in the images” (ibid., 431b) through abstraction. For instance, when thinking of a material thing, like a ‘snub nose’, the mind can employ this image to think of an abstract thing, like ‘curvature’. Thus, the mind “thinks of the curve apart from the flesh in which it exists” (DA.431b). So, just like intelligibles always exist in
sensibles, thought always employs images. Still, thoughts themselves are not images, but rather abstract reasonings which are derived from and illustrated by concrete images. This
25
explains why Aristotle recognizes both sensation and reason as sources of concepts: the former is necessary to provide images, but only the latter can extract universal judgments from these images through abstraction (A-15).
Conclusion
Aristotelian “simple thoughts” are mental objects (A-1) involved in thought and understanding (A-2). They are representations of something other than themselves (A-3), and resemble the things they represent (A-4). They are elementary entities resembling individual words (A-5), and they function as constituents of more complex mental acts (A-6).
Furthermore, like words by themselves, they do not posit the existence of these objects but only describe their nature, and thus do not have a truth-value (A-7). Specifically, they represent essences, which are real entities existing outside the mind (A-8). Because of this, they are also general representations which can apply to many individuals (A-9). Such representations are acquired through sensation (A-10) and stored in memory (A-11), and in their turn make knowledge possible (A-12). Thoughts and mental images are closely related, and the former never occur without the latter (A-13). However, while images are always of things which can be pictured, thoughts can be of many things, some of which are not sensible (A-14). While sensation is necessary to provide images, only reason can extract universal judgments from these images through abstraction (A-15). Aristotelian thoughts share most of these properties with modern concepts, except for their relation to real essences and to mental images.
26
Chapter 2. Concepts in Hellenistic philosophy
Although Plato and Aristotle usually get the most attention in discussions of classical contributions to the philosophy of cognition, Hellenistic philosophers have much to offer for the present study. Unfortunately, most of their original works have been lost. Still, there are sources which allow reconstructing their views, such as citations in the works of other writers, and summaries of their theories by authors such as Cicero (106-43 BC), Aetius (fl. 1st or 2nd century AD), Clement of Alexandria (150-210 AD), Sextus Empiricus (160-210), and
Diogenes Laertius (fl. 3rd century AD). On the basis of such extant sources, I discuss the contributions of Epicurus and the Stoics to the theory of concepts, and argue that their works offer a sustained and consistent account of the role of mental objects in cognition which often closely resembles the modern theory of concepts.
1. Epicurus
Born some 50 years after Aristotle, Epicurus was most strongly influenced by the Pre-Socratic materialism of Democritus. Like him, he held that everything which exists is matter: “the whole of being consists of bodies and space”, and beyond these, there is “nothing which by mental apprehension or on its analogy we can conceive to exist” (DL. 10.37). This theory held considerable explanatory power, and was applied by Epicurus not only to physics, but to psychology and epistemology as well. As will be shown, it also serves as a foundation of Epicurus’ theory of concepts. In the following, I first show that Epicurus recognized various kinds of mental objects, including ‘notions’ and ‘preconceptions’. Then, I discuss the role of preconceptions as standards of knowledge and as elements of thoughts. After this, I turn to Epicurus’ theory of the origin of preconceptions in sensation, and explain how this grounds his theory of the infallibility of preconceptions. Finally, I argue that Epicurus’ recognition of preconceptions of general kinds shows that he did actually recognize a role for reason in the formation of concepts, albeit a limited one.
Mental objects in Epicurus
Epicurus does not share his technical vocabulary with Plato and Aristotle. He only uses noema and ennoia, the terms used by these philosophers for ‘thought’ and ‘notion’, in passing. For instance, he mentions the "notion [ἐννοίᾳ] of infinity" (DL. 10.57) and the "notion [ἔννοιαν] of the body as a whole" (ibid., 69), and sometimes he uses the expression "quick as thought [νοήματι]" (ibid., 48). Instead, Epicurus prefers the terms prolepsis
27
[πρόληψις], which is usually translated as ‘preconception’, and rarely also epinoia [ἐπίνοια] and ennoema [ἐννόημα], both translated as ‘notion’.
Modern authors commonly refer to ‘preconceptions’ as ‘general concepts’ (cf. Bailey 1928: 424, Long 1971: 121, Rist 1972: 3), and identify them with other terms translated in this way, such as ennoema and epinoia (cf. Sedley 1973: 14-15, Nemeth 2017: 65). Epicurus' ancient interpreters equated them with the ‘thoughts’ and ‘notions’ of Aristotle and Plato as well. For instance, Diogenes Laertius states that a preconception is “a sort of apprehension [κατάληψιν] or right opinion [δόξαν ὀρθὴν] or notion [ἔννοιαν] or universal thought
[καθολικὴν νόησιν] stored in the mind” (DL. 10.33). Likewise, Clement of Alexandria defines preconception as “a belief [πίστιν] of the intellect, a grasp [ἐπιβολήν] of something evident, or an evident conception of a thing [ἐπίνοιαν πράγματος]” (Cl. Str. 2.4.16). Cicero also identifies preconceptions with notions, explaining that he “calls a notion [notio] what the Greeks call
prolepsis or ennoia” (Cic. Top. 7.30). From such passages, it appears that preconceptions are
the same as notions, which means they can be identified as mental objects (E-1). But though the difference between the terms ‘preconception’ and ‘notion’ might seem merely
terminological, Epicurus’ use of the former is in fact important: as will be shown, it reflects his thoroughly empiricist theory of concept formation. But before turning to this topic, I will first show why preconceptions are so important for Epicurus’ psychology and epistemology.
Preconceptions as standards of truth
In ancient accounts, preconceptions are mostly discussed in the context of Epicurus’ theory of knowledge. This theory claims that “our sensations [αἰσθήσεις] and preconceptions [προλήψεις] and our feelings [πάθη] are the standards of truth” (DL. 10.31). As was shown in the previous chapter, Aristotle held that uncombined thoughts cannot be false. Just like modern concepts, they do not involve any claim of real existence by themselves. But Epicurus’ preconceptions are not only never false: they are always true. In one passage, Epicurus claims that “preconceptions [προλήψεις] are always clear [ἐναργεῖς]” (ibid., 33). In another, “true preconceptions” are contrasted with “false assumptions” (ibid., 124). Thus, unlike Aristotle’s simple thoughts and modern concepts, preconceptions are always true (E-2).
The reason why preconception are standards of truth is indicated by their name: a preconception, or pro-lepsis, is literally “a ‘grasp’ that has been obtained before an inquiry” (Asmis 1984: 22). And indeed, according to Epicurus, without a preconception “nothing can be understood, sought, or debated” (Cic. ND. 1.16, cf. SE. 10.21). Thus, making judgments
28
requires having preconceptions. For instance, to make the judgment “the object standing yonder is a horse or a cow” requires “knowing by preconception the shape of a horse or a cow” (DL. 10.33). The need for previous possession of preconceptions extends even to the use of words: “the primary notion [πρῶτον ἐννόημα] of every term employed must be clearly seen, and ought to need no proving” (ibid., 38). Similarly, he states that if “the word ‘man’ is uttered, we immediately think [νοεῖται] of his shape by a preconception” (ibid., 33). This makes clear that preconceptions are ‘standards of truth’ because without them no knowledge would even be possible (E-3). Furthermore, they belong to thought, for we ‘think’ of a shape ‘by’ a preconception (E-4). More specifically, they are mental representations, as their function in thought is to represent objects such as men and horses (E-5). Because they
represent such objects to the mind, they allow distinguishing objects (E-6) and making correct judgments about their nature and their kind (E-7). Finally, they are connected to single words or expressions, each of which has a corresponding notion (E-8). Preconceptions share these properties with both the simple thoughts posited by Aristotle, and the concepts of modern philosophy. All three of them occupy the same position between language and reality, and all are required to make any knowledge and thought possible at all.
As was mentioned, another reason why preconceptions are standards of truth is because they are always true, and thus constitute a reliable source of knowledge (cf. Long & Sedley 1987: 89). According to Epicurus, preconceptions are derived directly from sensations, which themselves are always true depictions of real objects (cf. Theoph. DS. 1.7, DL 10.38, 46, 48). Thus, he claims that “all our notions [ἐπίνοιαι] are derived from perceptions
[αἰσθήσεων], either by actual contact or by analogy, or resemblance, or composition, with some slight aid from reasoning” (DL. 10.32), using the term ‘notion’ or epinoia which is generally regarded as a synonym for prolepsis. And indeed, in the same passage Epicurus describes a preconception as a “recollection of an external object often presented” (ibid., 33). The same link between preconceptions and stored sensations appears in a work by Cicero, where Lucullus the Epicurean argues that the mind “seizes on some impressions for its immediate use, while storing away others as the source of memory” (Cic. Ac. 2.30). In conclusion, as sensation is always of a real object and preconception is directly derived from it, there is no room left for error in either of them, and both are guaranteed to be true. This means that unlike modern concepts, many preconceptions are directly and automatically derived from sensation (E-9). For this reason, they seem to be mental images rather than abstract items (E-10). Still, the mind may use sensory data to construct new notions by combining them or altering them in certain ways (E-11). Furthermore, the identification of
