Chapter 6 - Concepts and Concept Formation in Aristotle’s Logical Works

Summary

Problematisation about concepts and related notions in Aristotle greatly depends on the particular work or works under consideration. Richard McKirahan’s contribution ‘Concepts and Concept Formation in Aristotle’s Logical Works’ argues that an account of how concepts are formed may be extracted from passages of the Organon, on topics relating to the problem of how we gain knowledge of scientific principles, in combination with Aristotle’s statements on the relations among objects in the world, affections in the soul, and language. Aristotle’s view is pieced together on the basis of the account in Posterior Analytics 2.19, supplemented by the parallel account in Metaph. A.1, as well as Aristotle’s remarks on the utility of dialectic for the sciences in Top. 1.2 and the brief discussions of epagōgē (frequently and controversially rendered by ‘induction’: An. Post. 1.18 and 1.31). The chapter also addresses questions about the nature of perception as conceived in the Organon, the epistemic status of experience, the transition from the awareness of individuals to the grasp of universals, the status of nous, and the relation between universals and concepts.

1 Introduction

In this paper I distinguish between two notions of concepts: (a) an individual’s concept of X and (b) the concept of X. (a) If I have a concept of X, I will typically be able to recognize X’s and distinguish them from things that are not X’s, and I will be able to say a number of things about X’s. My concept of X may be different from yours. For example, we might disagree about whether something is an x. Even if we agree, one of us might be unable to distinguish different kinds of X’s (to recognize something as a snake but not to know what kind of snake it is). And it may well happen that some of the things someone says about X’s are false. (b) When we speak of the concept of X, we are no longer talking about individual people’s understandings of X’s, but the correct concept of X. In Aristotelian terms having this concept of X this will include knowing the essence of X’s and as I shall argue, much more about X as well.

Although the vocabulary of concepts is barely present in Aristotle’s logical works, the second notion of concept is centrally important in the theory of demonstrative science presented in the Posterior Analytics. In the first place, ‘that of which there is unqualified epistēmē (scientific knowledge) cannot be other than it is’ (An. post. 71b15–16), that is, it is true and necessarily so. This means that the conclusion of any scientific demonstration is true and, given Aristotle’s notion of demonstration, this means that it is the deductive consequence of the premises employed in the demonstration and that the premises are true as well. The premises are immediately or ultimately derived from the principles of the science, which for the same reasons are also true. The principles of a science include definitions of all the terms (subjects and attributes) employed in the science’s demonstrations. The fundamental question of how we acquire (undemonstrated) knowledge of principles is addressed only in the final chapter of the work. The account given there largely consists in identifying a progression of mental states (hexeis) from aisthēsis (sensation/perception) to memory to empeiria (‘experience’) and a grasp of universals, to the grasp of principles (which Aristotle calls nous). The account is incomplete and unclear and problematic in many ways. Fortunately, other Aristotelian works make it possible to solve some of the problems, to fill some of the gaps, and to provide a reasonably full and coherent Aristotelian account of the contents of these mental states and the kind of grasp each of the states provides of its contents. At the same time there emerges an account of concept acquisition that covers both kinds of concepts outlined above: moving from an individual’s acquisition of concepts to the (scientific) concept in all its richness.

In what follows, after a brief defense of the view that (at least in the Organon) ‘concept’ is a reasonable translation of noēma (Section 2), I devote the lion’s share of this paper to a detailed reading of relevant passages in the Posterior Analytics (2.19) and Metaphysics (A.1) which I read as an empiricist account of concept acquisition that includes concepts of both universals and particulars (Section 3). I argue that these two passages present a consistent picture, addressing the topic from different viewpoints. However, unlike earlier interpretations including my own (in McKirahan 1992) I now float the possibility that the Metaphysics account improves on the Posterior Analytics account in places and so may be read as a revised version rather than a mere rewording.

Section 4 sketches an account of how Aristotle might have conceived the process of going from knowledge of (universal) concepts to scientific principles, in effect from conceiving something as a general truth to conceiving it as a fundamental truth. Section 5 sketches how the results of Section 4 may be relevant grosso modo to sciences as we know them and setting an intelligible (if hardly obtainable) goal for science. Finally, Section 6 returns to the notion of concepts in Aristotle and clarifies the ways in which the discussion in this essay may or may not be related to Aristotle’s thinking in the Posterior Analytics.

2 Concepts in the Organon

The first question that must be faced is whether concepts make any appearance at all in Aristotle’s logical works. It is commonly assumed that they do and that they are to be found in the account of how we obtain knowledge of scientific principles in Posterior Analytics 2.19.¹ They are taken to be universals – the universals that have come to rest in the soul (An. post. 100a6–7), a condition of the soul that is somehow linked to empeiria, as either (a) something that empeiria includes or (b) a more advanced condition of the soul between empeiria (which on this account works at the level of particulars) and nous (the intellectual grasp that we have of scientific principles).² But Aristotle does not say that when a universal, say [dog]³ has come to rest in a soul, that what’s in the soul is a concept of dog or the concept of dog. The vocabulary is not psychological (except for psukhē) but ontological and logical. In the first part of his treatment, he calls aisthēsis (which covers what we call both sensation and perception) a dunamis sumphutos kritikē (congenital capacity of [or for] discerning) (99b35) and he distinguishes animals in which ‘there occurs a persistence of the percept’ (aisthēma) from those in which it does not occur. But there is no word corresponding to ‘percept’ when he discusses universals. That is, he does not say anything that could be taken as a statement that universals come to rest in the soul when there occurs a persistence of the concept.

In the parallel passage in Metaphysics A.1, which fills some of the lacunae in the An. post. account,⁴ Aristotle speaks of the ennoēmata of empeiria, declaring that ‘from many ennoēmata of empeiria there comes to be a single universal hupolēpsis about the similar things’ (Metaph. A.1 980b5–7).⁵ He immediately goes on to contrast ‘the possession of an hupolēpsis that this helped Callias when he was suffering from this particular disease,’ (which is a matter of empeiria), with the hupolēpsis ‘that it has helped everyone of a particular definite sort when suffering from this particular disease’ (which is a matter of the medical art/craft (tekhnē)), where the disease in question is specified and distinguished from other diseases. Ennoēma is a psychological term that can be translated as ‘concept’ or ‘thought’ as designating a kind of mental content, as in the hupolēpsis concerning Callias. The word hupolēpsis has a range of meanings (‘judgment,’ ‘supposition,’ ‘assumption,’ ‘opinion,’ ‘belief’), all of them epistemological. Aristotle himself identifies epistēmē, doxa and phronēsis and their opposites as kinds of hupolēpsis (De an. 3.8, 427b25–6). Evidently these hupolēpseis have propositional content. Do ennoēmata have propositional content, too? For example, can I have (1) an ennoēma that Socrates is bald, or only (2) ennoēmata of Socrates and bald? Can I have an ennoēma (3) of human being, or only (4) of a particular human being?

This slender amount of information can be supplemented by the usage of a word related to ennoēma that occurs in another of Aristotle’s logical works, which points to answers to these four questions. The passage goes as follows:

(A) Spoken sounds correspond to (are sumbola of) affections (pathēmata) in the soul, and written marks correspond to spoken sounds. And just as written marks are not the same for all people, neither are spoken sounds. But what these⁶ are in the first place signs of – affections (pathēmata) of the soul – are the same for all; and what these affections are likenesses (homoiōmata) of – actual things (pragmata) – are also the same. These matters have been discussed in our work on the soul and do not belong to the present subject. (B) Just as there can be noēmata in the soul that are neither true nor false while some noēmata are necessarily one or the other, so also with spoken sounds. For falsity and truth have to do with combination and separation. Thus names and verbs by themselves – for instance, ‘human being’ or ‘white’ when nothing further is added – are like a noēma that is without combination or separation; for so far they are neither true nor false.

(Int. 16a3–16, translation adapted from Ackrill)

This establishes that some noēmata (those that are true or false) have something corresponding to propositional content while others (those without combination or separation) do not. If noēmata are the same as ennoēmata, as I suppose,⁷ then questions (1) and (2) have affirmative answers: some ennoēmata can have propositional content and others do not. In Section 4 I call universal noēmata without propositional content A universals and those with propositional content B universals. Questions (3) and (4) have affirmative answers too: the use of ‘human being’ apparently as an example of a noēma in the De Interpretatione passage cited gives good reason to suppose that there are ennoēmata of universals, while the interpretation of empeiria that will be developed below entails that there are ennoēmata of particulars as well.⁸

It is not clear that the passage quoted above refers to concepts. The pathēmata are ‘the same for all,’ which I understand to mean ‘for all people.’ But as was noted above, different people may have different concepts of the same thing, as Aristotle well knows.⁹ If this is so, pathēmata are not concepts. If, as Ackrill thinks,¹⁰ the noēmata of section B above are the same as the pathēmata of section A, then concepts do not appear here at all.

The other possibility is that the noēmata referred to are not the same as the pathēmata. In fact, the two words have different associations. A pathēma is something that is experienced – in this case, by the senses, while a noēma is something that results from thought. We both hear a loud noise (we have the same pathēma) but your resulting noēma may be different from mine. You may take it to be a sound made by a train, and I may take it to be the sound made by a factory whistle. On this reading a reasonably strong case can be made for identifying noēmata with concepts.

This reading finds support in Ackrill’s translation of noēma as ‘thought’ – by which he means, I take it, not an act of thinking but the thought we are thinking when we are thinking – roughly speaking, the contents of an act of thinking or a mental representation of the object of thought. In the rest of this paper I will suppose that Aristotle’s noēmata are concepts.

Aristotle’s remarks on the relations among reality, the mind, and language are unclear and leave important questions unanswered. (1) How do noēmata of universals come about in the soul, supposing that they are grounded in perception of particulars? (2) Are the noēmata of universals the same for all? (3) Even if we agree with Aristotle that ‘in fact the soul is such that it can be affected in this way’ (An. post. 100a13–14) we must feel that without offering some account of how this happens there are serious gaps in his accounts of perception, thought, knowledge in general, and scientific knowledge in particular. Further, regarding scientific knowledge, we are left wondering how – even if he gives an account of how we reach universal terms and universals with propositional content – he might account for our knowledge that certain terms and propositions are related to one another in such a way that taken together they constitute a science – in particular, how he might account for coming to know some of the universal propositions as principles, the most basic propositions in the science: in Aristotle’s terms, how we come to have nous as well as epistēmē. In the remainder of this paper, I aim to present a reading of relevant passages of the Posterior Analytics and the Metaphysics that goes some way towards closing some of these gaps.

3 Aristotle’s Developmental Accounts

Attempts to discover Aristotle’s views on concept acquisition focus on two passages,¹¹ from two different works: the last chapter of the Posterior Analytics (2.19) and the first chapter of the Metaphysics (A.1).¹² 2.19 takes up two questions in connection with scientific principles: ‘(1) how do they come to be known and (2) what is the state (hexis) that recognizes them?’ (99b17–19). We do not have scientific knowledge (epistēmē) of them because scientific knowledge is a matter of knowing through demonstration and since as principles they are by definition indemonstrable (99b20–22), we must know them in a different way. Aristotle summarily rejects as odd (atopon) the Platonic view that they are simply inherent (enousai) (99b25–27) and proceeds to take up question (1) (99b28–100b5) and then question (2) (100b5–17). Our focus will be on the answer he gives to question (1). A.1 is devoted to a very different question: What is wisdom (sophia), the supreme kind of knowledge? A.1 concludes that sophia is knowledge of certain principles and causes (982a2–3) and in working towards this conclusion it gives an account that is similar to that in 2.19, but fuller than 2.19 at crucial points.¹³ Since neither passage occurs in a context devoted to concept acquisition, it is not surprising that Aristotle does not address this topic explicitly even if some of what he says is relevant.

Both accounts begin with aisthēsis, which as we have seen 2.19 calls ‘a congenital capacity of (or for) discerning’ (99b35), and which A.1 says makes us aware of differences/distinctions (diaphorai) (980a27).¹⁴ I take it that the point of these passages is essentially the same. Little is said in either text about aisthēsis. Both move on immediately to memory, described as ‘a persistence of the percept (monē tou aisthēmatos)’ (99b36–37) a capacity found in some animals but not in others. ‘In animals where it is not found (whether it does not occur at all or concerning the things for which it does not occur) there is no knowledge (gnōsis) apart from sensing/perceiving (aisthanesthai)’ (99b37–39). A.1 in its search for the highest kind of knowledge observes that creatures possessing memory are more intelligent and better able to learn than those lacking it. It then goes on to say that unlike human beings, who possess tekhnē and reasoning (logismos), the animals that possess memory live by their phantasiai (appearances, images, impressions) and memories. I take it that the phantasiai in question are sensory images and since they are different from memories, I take it that they are fleeting sensory images, equivalent to the percepts of 99b37. We are aware of them when they are present, but once gone they are gone, leaving no traces. Without memory, then, an animal cannot grasp that what it is perceiving on one occasion (say a particular shade of green) is the same or different from what it has perceived on another occasion; in fact, it cannot even grasp that there have been other occasions. And perhaps it is not able to perceive it as green.

In some animals the percepts persist. How this happens is not Aristotle’s concern in these passages, so it is not surprising that he says nothing about it here. The brief account in De Memoria 1,¹⁵ however, goes some way to fill the gap consistently with his remarks in 2.19 and A.1. Aisthēsis generates a process of movement that stamps an impression (tupos) of the percept (aisthēma) in the soul (450a31). Depending on the condition of the soul the impression may not last (in which case the person or animal has a bad memory) or may not be formed at all (in which case memory does not arise) (450a31–b11).

When we remember something, we remember some thing; when we see something and remember it, we remember it as the same thing. This is due to the persistence of the percept. Or rather percepts, since Aristotle speaks of ‘many occurrences of this kind taking place’ (100a1) and ‘memory occurring many times of the same thing’ (100a4) as prerequisite for reaching the stage of empeiria. Our memory of something is typically not simply the result of a single imprint that is repeated many times, but a composite of different imprints. Thus, one’s memory of something can be more or less detailed, more or less clear, more or less rich depending in part on the frequency, the intensity and the variety of circumstances in which we come into sensory contact with it. In this sense memory can grow. Aristotle does not emphasize this aspect of memory in either of our two texts, but it is implicit in the account of the advance from memory to empeiria.

At this point (100a1–3, 980b25–28) both accounts continue to refer to the scale of cognitive capacities in animals. The rather vague assertion in 2.19 that ‘When many occurrences of this kind take place, a certain difference then occurs, with the result that for some from the persistence (monē) of such things there comes to be logos and for others there does not’ (99b36–100a1) – where ‘the persistence of such things’ must refer to memory and ‘such things’ must refer to aisthēmata (cf. 99b36–100a1), and logos must refer to one or more of the higher capacities (empeiria, tekhnē, epistēmē, nous) – is replaced in A.1 by the more precise statement that ‘the other (sc. animals) live by their phantasiai and memories and have but a small share of empeiria, while human beings live by tekhnē and reasonings (logismoi) as well’ (980b26–28).

In A.1 animals are granted a small share of empeiria (although no explanation of this claim is given), while tekhnē (and clearly also epistēmē) and reasoning are the exclusive province of human beings. It is tempting to suppose that Aristotle wrote this part of A.1 with 2.19 before him and decided that logos (a word found shortly earlier in the text of the Posterior Analytics (97a19, b13; 98a12, 99a3, a25) where recent translations render it variously as ‘definition,’¹⁶ ‘formula,’¹⁷ ‘rule,’¹⁸ ‘argument’¹⁹ and ‘account’²⁰) was not clear enough, and that, having for whatever reason decided to allow animals a share of empeiria, and so being forced to relocate human superiority higher up, located it in the possession of tekhnē (which presupposes empeiria) and reasoning, which (as the very word he uses, logismos, indicates) presupposes the use of the uniquely human achievement that is language (another meaning of logos).

The following sentences in 2.19 and A.1 show the same pattern.

So from perception memory comes about, as we say, and from memory of the same thing occurring many times comes experience, for memories many in number are one experience.

(An. post. 2.19, 100a3–6)

From memory experience comes about in human beings, for many memories of the same thing produce the capacity of one experience.

(Metaph. A.1, 980b28–981a1)

Like 2.19, A.1 specifies that the many memories that produce experience are memories of the same thing, but more carefully than 2.19, A.1 does not simply identify experience with the many memories but indicates that the many memories generate the experience. A.1 (but not 2.19) now adds (981a1–5) that empeiria is thought to be quite similar to tekhnē and epistēmē and insists that the latter two states are different from one another, and are attained, if at all, through empeiria, quoting to good effect a bon mot of Gorgias’ pupil Polus: ‘experience makes art (tekhnē), but inexperience makes luck.’²¹

How to understand the following assertion of 2.19 has for some time been a source of disagreement.

From experience, or when the entire universal has come to be at rest in the soul – the one beside the many, whatever is one and the same in them all – the starting-point of art and science comes about: the starting-point of art if it is concerned with coming to be, but the starting-point of science if it is concerned with what is.

(An. post. 2.19, 100a6–9)

The dispute depends on the word ‘or’. There are three ways of taking the claim: (1) ‘or’ means ‘i.e.’: experience is identical with the state where the entire universal has come to be at rest in the soul;²² (2) ‘or’ has a corrective or progressive sense and means ‘or rather’ (cf. French ou bien): experience is the state prior to the state where the entire universal has come to be at rest in the soul;²³ (3) ‘or’ means ‘or’ – experience and the state where the entire universal has come to be at rest in the soul are two different routes to the starting point of art and science.

As far as I know only one person²⁴ has favoured interpretation (3). I think it obvious that Aristotle is not describing two different ways of reaching the principle of art and science. Interpretations (1) and (2) have supporters. The corresponding passage in A.1 is clearer on the subject and to my mind settles the question in favour of (2).²⁵

Tekhnē arises when from many ennoēmata of empeiria there comes a single universal hupolēpsis concerning similar things. For it is a matter of empeiria to have a grasp that this helped Callias when he had this illness, and also Socrates and so for many individuals. But it is a matter of tekhnē to grasp that it helped all people of a certain kind, marked off in a class, who had this illness, for example, phlegmatic or bilious people burning with fever. For as far as performance goes, empeiria appears no different from tekhnē; in fact we see people with empeiria who are even more successful than those who have a theoretical knowledge of the subject but lack experience. The reason is that empeiria is knowledge of the particulars whereas tekhnē is knowledge of the universals, and all actions and instances of coming into being are concerned with the particular.

(Metaph. A.1, 981a5–17)

A.1 (especially 981a15–16) could not make it clearer that empeiria stays at the level of particulars (Callias, Socrates) and that the universal hupolēpsis comes from (ek) the ennoēmata that accompany empeiria just as empeiria comes about from (ek) memory and memory comes about from (ex) aisthēsis. Once again, A.1 clarifies an ambiguous statement of 2.19, and this time it provides an example to illustrate the difference. The ‘similar things’ are Callias, Socrates, and the other individuals who ‘had this illness,’ and the universal is ‘all phlegmatic people burning with fever’ or ‘all bilious people burning with fever.’ ‘All’ such people, because the universal applies to all people of a certain kind, are marked off in a class. And the illness is marked off too. No longer just ‘this illness’ but ‘burning with fever.’ At the level of empeiria we have not yet got beyond the level of dealing with individual cases. On the basis of our experience, we are able to spot similarities and to prescribe a cure, but unable to state precisely what similarities Callias and Socrates have that make them prone to this illness and why precisely such people are helped by ‘this.’ We can do these latter things when we have mastered the tekhnē, of medicine. We have theoretical knowledge (‘phlegmatic’ and ‘bilious’ are technical terms in the Hippocratic theory of the four humours²⁶) that enables us to classify people into medically relevant types and to prescribe treatment accordingly for each type of people. And similarly for illnesses, although Aristotle’s example does not point this out. There are many illnesses that result in a high temperature and a doctor will be trained to recognize (perhaps from other symptoms as well) not only that the person is feverish but precisely what disease the fever indicates. By contrast, at the level of experience we may indeed be able to prescribe the correct cure, but without being able to explain why it is correct apart from saying that has helped Socrates, Callias and many others in the past. This makes it clear that the description of empeiria as the product of ‘many memories of the same thing’ (980b39, also less clearly 1000b3–4) can mean not only memories of Socrates being cured by this, but also memories of Callias and others too being cured by this. In this case the ‘thing’ in question is that this cured a number of people who are alike in some unspecified way. The word dunamis (‘capacity’) found in the A.1 account (981a1) is an important qualification as well and can be viewed as correction of the 2.19 account. It points to the fact that simply remembering several similar cases can but need not bring about the state of empeiria.

Just as a person’s memory of something may grow over time, as we become better acquainted with it through our senses, our empeiria of a phenomenon can be expected to grow too. Some craftspersons have more experience than others and as a result can do their work better. The nature of this ‘growing experience’ is well discussed by Thomas Johansen in this volume.

At this point A.1 and 2.19 diverge. A.1 continues to make the case that sophia consists in knowing certain principles and causes, a state achieved with epistēmē but not yet attained at the stage of empeiria, while 2.19 continues its account of how we achieve knowledge of scientific principles. From here on only only the 2.19 account is relevant to our present purpose.

What follows (down to 100b5) consists of several parts. The first (A: 100a6–11) completes the developmental account and asserts that the developmental account solves the problem stated at 99b25–26: knowledge of scientific principles comes from aisthēsis. The second (B: 100a12–14) presents a simile of routed soldiers which Aristotle says illustrates the relevant capacity of the soul. The third (C: 100a14–b3) offers what he curiously describes as a clearer account of what he has just said. The fourth (D: 100b3–5) repeats the conclusion stated at 100a10–11 and declares that the way we come to know ‘the first things’ is ‘the same way that aisthēsis imparts (empoiei) the universal.’ These difficult and disputed seventeen lines require careful treatment.

Section A

…(i) when the entire universal has come to be at rest in (en) the soul – the one beside (para) the many, whatever is one and the same in (en) them all – the principle of art and science comes about:²⁷ (ii) the principle of art if it is concerned with coming to be; the principle of science if it is concerned with what is. (iii) So the states are not present in us as determinate states, nor do they arise from still other, higher cognitive states, but from aisthēsis.

(100a6–11)

This section has three parts. Part (i) gives rise to some questions. (a) What does Aristotle mean by the entire universal? See below for a suggestion.²⁸

(b) In what way is the universal in (en) the soul both beside (para) the many particulars and in (en) them, and does this claim reveal anything about the relation between universals and concepts? The former locution is used of the universal ‘at rest in the soul’ as opposed to the universal ‘in the particulars.’ We start with the familiar notion of universals being ‘in’ particulars and now in addition there it is in the soul as well, but not ‘in’ the soul in the same way as it is in the particulars. I suppose that Aristotle means that the noēma of human being is different from the noēmata of Socrates and other individual human beings, just as the universal [human being] is different from Socrates and other individual human beings. The two cases are different, however, in that the universal does not exist separately from the individuals, whereas the noēma of human being does exist separately from the noēmata of Socrates and other human beings. Aristotle does not elaborate further on this matter. However, the preposition para is a good choice for expressing the difference, since in one of its uses it can be translated ‘aside from,’ ‘in addition to,’ ‘alongside,’ and ‘over and above.’ This might be a way to distinguish the concept of X from individual X’s: the concept of X is something over and above the individual X’s and our knowledge of it is grounded in our familiarity with the individual x’s, while our knowledge of individual X’s as X’s is grounded in our knowledge of the concept.²⁹ The concept of X will not have any of the peculiarities of individual X’s but will in some sense contain what all the X’s have in common as X’s. Aristotle does not provide further information on the meaning of para.

(c) How does a scientific principle come about? Aristotle is obviously talking not about what happens when principles (which like all scientific facts are necessary and eternal) come into existence, but when they come about (come to be) in us, that is, in our soul. Being universals, they come to be in the soul in the same way other universals do. However, for Aristotle scientific principles are not just any old universals in our souls, and we may feel disappointed by an answer that accounts for all universals, including universals in non-scientific as well as in scientific contexts.

In mitigation of this last charge, it is fair to point out that Aristotle can claim to have answered the problem he set himself at 99b18 (How do scientific principles come to be known?) given the context in which he set it (finding an answer that does not depend on pre-existent knowledge). All scientific facts are universal, both the principles and their deductive consequences. By showing that it is through aisthēsis, memory and empeiria that we come to know universals he has answered how we come to know universal scientific facts. It is just that he has not said anything about how we know that a universal fact is a scientific fact and how we know among scientific facts which of them are principles. That is, unless he is pointing in this direction by speaking of ‘the entire universal.’³⁰

Part (ii) is straightforward: tekhnē deals with coming to be and epistēmē with what is. The crafts Aristotle has in mind, such as carpentry, have products, they are involved with making things, whereas epistēmē is theoretical, concerned with knowledge, not with action or production. This distinction is found elsewhere as well.³¹

Part (iii) is also straightforward except for its mention of ‘the states (hexeis).’ Are they epistēmē and knowledge of scientific principles (later identified as nous), or epistēmē and tekhnē, or all three? According to the developmental account, all three states develop out of aisthēsis. On the one hand, it might be urged that only nous is intended, since things known by epistēmē are known as a result of demonstrations from principles which depend on nous, and that something analogous holds for tekhnē. But I think it is meant to hold for all three states, since the present discussion does not treat demonstrations but knowledge of the scientific or technical (tekhnē-related) facts some of which will turn out to be principles and some of which will not.

Section B

… as in a battle after a rout has occurred when one man comes to a stand, another does too and then another – until it reaches a starting-point.³² And in fact the soul is such that it can be affected in this way.

(100a12–14)

This simile likens the state in which ‘the entire universal has come to be at rest in the soul’ (100a6–7) to a situation on a battlefield where one soldier after another ‘comes to a stand.’ The general idea is simple. Soldiers who have been routed are no longer behaving like soldiers, working together as an army in an orderly fashion; they are behaving like individuals with no discernible order. To perform their work as soldiers they need to be assembled in formation. Likewise, the individual sense impressions need to be associated in appropriate ways to form a single memory and individual memories have to be associated in different appropriate ways to form a single empeiria, and many empeiriai have to be associated in still different appropriate ways for a universal to come to be at rest in the soul. This much is straightforward.

The problem is that it does not account for epistēmē, let alone nous, and we will not have forgotten that 2.19 declares at the outset that it will make clear how the principles (arkhai) become known (99b17–18). Another thorny problem is what to do with the word translated ‘starting-point’ – the very word Aristotle uses for scientific principles. It has been translated as ‘original position,’³³ ‘original formation,’³⁴ ‘beginning,’³⁵ and ‘a position of strength’³⁶ – in all cases unconvincingly, since none of these renderings except for ‘a position of strength’ makes sense if one considers the plight of routed hoplites who could not return to their original position or even re-group elsewhere into a phalanx formation while being pursued and slaughtered by the spear-throwing enemy hordes. Barnes’s translation does make sense but even so, it is objectionable for two reasons. First, it is a translation not of Aristotle’s text as transmitted in the manuscripts, but of an emendation Barnes introduced in despair of making sense of the received text.³⁷ An interpretation of the received text that gives good sense is obviously to be preferred. Second, Barnes’s translation of the emended text is forced. The word in question, alkē, means ‘strength’ or ‘force,’ not ‘a position of strength’ as Barnes renders it.³⁸

What I believe is the correct solution to this problem was first discovered (but not adopted) by myself³⁹ and was first adopted by Lesher,⁴⁰ namely, that the simile ends after ‘and then another’ and the remaining words ‘until it reaches a starting-point/principle’ refer not to the soldiers but to the end of the developmental process from awareness of particulars by means of aisthēsis to the grasp of principles by nous. What reaches the principle is the process itself, or perhaps more likely (in view of the immediately following words) what reaches the principle is the soul itself.

The passage is silent on an obvious question that arises at this point: if the soul is capable of being affected in this way, that is, if it can undergo the developmental process, how does it do it? Does it do it deliberately or is it simply the passive location in which the process takes place? I return to this below.

Section C

But let us say again what was said just now but not clearly. When one of the undifferentiated things comes to a stand, the first universal in the soul – for in fact it [the soul] perceives the particular, but the perception is of the universal (of human being, for example, not of Callias the human being). Once more it comes to a stand in these, until whenever the things that have no parts (ta amerē) and are universals come to a stand (for example, such and such a kind of animal, and then animal, and then likewise in this).

(100a14–b3)

What was said ‘just now’ had to do with ‘the entire universal’ coming to rest ‘in the soul’ (100a6–7) but what is said in section C has to do with moving from ‘the first’ (least general) universals (e.g., [human being]) to more general ones (e.g., [animal]), not failing to include any intervening universals (e.g., [mammal]). A reasonable way to avoid this inconsistency is to suppose that ‘what was said just now but not clearly’ is clarified not in all of the remainder of section C but only through the parenthetical remark. The first universal is in the soul when the perception is no longer just of the particular human being, but of the universal [human being]. The sequel then says that an analogous process leads to more general universals lodging in the soul. I will proceed on the basis of this reading.

The undifferentiated things (adiaphora) will be the particular instances of the universal. Of course Socrates can be differentiated from Callias as individuals (for example, they don’t look the same) but from the point of view of natural science they are the same: individual human beings or more specifically (perhaps for medical purposes) individual phlegmatic human beings with fever, and qua such they are treated identically. The word adiaphoros has this sense and application in its occurrence at 97b28–31,⁴¹ a passage near enough to the present context that section C reasonably counts as repeating what had recently been said. Granted that (as we are about to find out) there is a hierarchy of generality among universals, it is reasonable for Aristotle to refer to the universal so far obtained as the first universal.

The essential move in Aristotle’s view of what we might reasonably consider concept acquisition is found in the next claim, that the soul ‘perceives (aisthanetai) the particular but the perception (aisthēsis) is of the universal – for example, of human being, not of Callias the human being.’ Callias is an individual human being. We may have seen him and heard him many times and we may remember many things about him and may even have empeiria that is based in part or whole on these memories. But until we grasp the relevant universal ([human being]) we are not yet in a position to see him as a human being, or as Callias the human being, even though in fact he is a human being.⁴² After we acquire the universal [human being] we no longer need see him as Callias (who happens to be a human being whether we recognize it or not); for some purposes (for example, for the purposes of natural science) we may consider him simply as a human being. For other purposes we might see him as something else. For example, a doctor might regard him as a phlegmatic human being (who happens to be Callias) rather than Callias (who happens to be a phlegmatic human being). In my view, this process of coming to see a particular as a particular of a certain kind (as a human being, or as a phlegmatic human being) is what Aristotle calls epagōgē, which I discuss below.

The next sentence sketches how more general universals are acquired. First we acquire a number of ‘first universals’ and then notice that they naturally fall into groups, some more inclusive than others, and so we might move from ‘first universals’ such as human being, lizard, beetle and trout to higher universals such as mammal, reptile, insect and fish, and from there to higher levels.⁴³ Aristotle gives only the briefest outline here, but he says much more in other works, from the earliest, such as the Topics, to his later works in biology. A nudge in this direction is enough for his present purposes.

The words ta amerē, translated ‘the things that have no parts’ have caused difficulties for this straightforward interpretation of the sentence. (1) In what sense do universals have no parts? Could the particular instances of a universal be its parts? Even if this is what he means, it is not an Aristotelian way to describe the relation between universals and particulars. (2) Even if he uses ‘parts’ in this sense, will not all universals (not just the most general ones) equally have parts? If so, then how can Aristotle be specifically describing the most general universals? And if he is not, then he cannot be describing the ascent from the ‘first’ (i.e., most specific) universals to the most general, which undermines the interpretation proposed above, which is the only one that seems to fit the context. (3) Since at the stage he is describing we have already reached the level of universals, what is the point of saying ‘the things that have no parts and are universals’?

I propose the following solutions to these puzzles. (1) Aristotle might reasonably say that both particulars and universals are without parts, but for different reasons. He does just this at Metaphysics M.8 1084b13–15: ‘The universal and the particular or the element are indivisible (adiairetos), but in different ways, one in formula and the other in time’ (trans. Ross). Next, whatever the relation between particulars and universals may be, Aristotle never describes the relation between particulars and universals as being that of part to whole. Rather, he speaks of the universal as ‘in’ or ‘said of’ particulars.

(2) On the other hand, he does speak of one universal being part of another, notably when he says ‘the elements in the formula which explains a thing are parts of the whole; this is why the genus is called a part of the species, though in another sense the species is part of the genus.’⁴⁴ A species is part of the genus because the genus is divided into its species. But the genus is part of each of its species because it figures into their essence and definition. It is in this latter sense that Aristotle can describe the highest (most general) universals as without parts. If, as Aristotle has in mind, definitions are statements of essence and have the form of genus plus one or more differentiae, there can be no definition of the highest genera, since there is no higher genus for them to belong to, and so in this sense they have no parts.

(3) On the other hand, since as we have seen he holds that particulars are without parts in a different way, in our present passage he needs to specify that the partless things he is talking about are universals, not particulars, which he does when he says ‘until whenever the things that have no parts and are universals come to a stand.’ This is the solution to problem (3).⁴⁵

Section D

So it is clear that it is necessary for us to come to know the primary things by epagōgē, since this is how aisthēsis produces the universal in us.

(100b3–5)

Aristotle exploits the broad term aisthēsis, which covers both sensation and perception, to make his point here. We begin by sensing particulars, and then progress through the stages of memory and empeiria and next, when we have acquired knowledge of the relevant universal, we perceive them as particulars of a certain kind. He even talks of ‘seeing’ universals in this latter sense, as at 77b31 (‘something is said of all this and this again is said of all something else… and one can as it were see these by thought’).⁴⁶

This is epagōgē. Aristotle uses this word and the related verb epagein to refer to the way we come to see particulars in this latter sense (as particulars of a certain demarcated kind) elsewhere in the Analytics, principally in An. pr. 2.21 and An. post. 1.1.⁴⁷ Importantly for our purposes, epagōgē is not only how we come to see Socrates as a human being, but also how we come to see human beings as animals. Also importantly, epagōgē is the basis not only for identifications (identifying Socrates as a human being, identifying human beings as animals) but also for certain predications (‘Socrates is a human being’ and ‘human beings are animals’).⁴⁸ Section D uses the term epagōgē for the process of moving up the ladder, from seeing Socrates as a human being to seeing him as an animal, and as seeing human beings as animals. This process, however it works, brings us from the more specific universals to the more general ones. By gaining familiarity with a variety of universals we can see how they are related, and we ultimately grasp the uniquely best way to classify them (along with the individuals that belong to the species) by genera and species. In this context, ‘primary things’ could be either what he calls ‘the first universals’ – that is, the most specific universals, like [human being] (100a16) or ‘the things that have no parts and are universals’ – that is, the highest, most general universals. The context favours the latter possibility.

4 From Universals to Scientific Principles

So far, we have extracted an account of how we come to have knowledge of universals, an account based on examples of epagōgē which result in concepts of universals like [human being] and [animal], which I will call A-universals but which is evidently intended to apply as well to universals like [aspirin cures headaches], which I will call B-universals. The immediate context of 2.19 demands this and it is justified by Aristotle’s account of noēmata that are true and false.⁴⁹ Further evidence that Aristotle recognized epagōgē that results in knowledge of B-universals is his reference to epaktikoi logoi,⁵⁰ which result in universal conclusions, which are, of course, propositions, not terms, such as his example of an epagōgē where the conclusion is that this figure in the semicircle has angles equal to two right angles.⁵¹

Even so, as an account of how we come to have nous it falls short, in five ways. (1) It does not explain how we move from A-universals to B-universals. (2) It says nothing to guarantee that my understanding of universals is the same as yours – for example, that my concept of human being is the same as yours. (3) Even supposing that my understanding of a universal is the same as yours, it does nothing to guarantee that we have the correct understanding of it – that we have grasped the concept of human being. (4) It gives as an example the acquisition of an A-universal, whereas science deals with B-universals. (5) It does nothing to distinguish how we come to know principles from how we come to know non-principles in a science.

These are important omissions which could leave the impression that Aristotle was in the end unable to complete his account of scientific knowledge. In this concluding part of my essay, I attempt to show that this is not the case and simultaneously to support the interpretation I favour, that for Aristotle scientific knowledge involves knowing the concepts treated by the science in a peculiarly strong way possessed only by experts. Material that supports this interpretation is found elsewhere in the Posterior Analytics and in other works of Aristotle, principally the Topics (an early work which he might have assumed was familiar to readers of the Posterior Analytics).

Problem (1) is not really a problem for Aristotle. By the time we can recognize different people as people, we have become aware that they share certain features, that they are alike in certain ways. We identify things as humans on the basis of whether or not they have these characteristics. By the time we reach the stage of empeiria our soul has become a storehouse stocked with memories not just of individual things but of individual things along with (some of) their attributes – not only a memory of John, but also a memory that John is tall⁵² and we can recognize similarities in things – that John and Thomas are both tall.

To judge from the first chapters of De Interpretatione, Aristotle appears to have found the move from facts (such as that humans walk and talk) to propositions ((most) humans walk’ and ‘(most) humans talk’) unproblematic. The relevant passages are these.⁵³

Spoken sounds correspond to affections of the soul and written marks correspond to spoken sounds.

(Int. 16a3–4)

Just as there can be concepts (noēmata) in the soul that are neither true nor false while some noēmata are necessarily one or the other, so also with spoken sounds. For falsity and truth have to do with combination and separation. Thus names and verbs by themselves – for instance, ‘human being’ or ‘white’ when nothing further is added – are like a noēma that is without combination or separation; for so far they are neither true nor false.

(Int. 16a9–16)

A sentence (logos) is a significant spoken sound some part of which is significant in separation as an expression, not as an affirmation.

(Int. 16b26–27)

A statement-making sentence (apophantikos logos) is [one] in which there is truth or falsity.

(Int. 17a2–3)

A simple statement (apophansis) is a significant spoken sound about whether something does or does not belong.

(Int. 17a23–24)

The simple statements that are important to our present purpose are kataphaseis (assertions/affirmations), which state that something belongs to something, and apophaseis (negative propositions/negations), which deny that something belongs to something (Int. 17a25–26).⁵⁴ It follows that for Aristotle we can move freely and unproblematically back and forth between (a) things and facts and (b) concepts (simple and complex noēmata).

Problem (2) finds its solution in the aim of the Posterior Analytics, to describe what an achieved science looks like. This is the successful case, where all the principles have been correctly identified and all the demonstrations have been found. In this ideal situation, experts – those who have mastered the science – know a great deal more about, say, human beings and animals than ordinary people who are familiar enough with these universals to be able to identify Socrates and Callias as human beings and to identify human beings and dogs as animals. Expert knowledge is far richer, enabling its possessors to understand the place of each entity and each fact in the complex network of deductive relations that constitutes a science. For example, expert knowledge of the Pythagorean theorem (Euclid, Elements 1.47) is not just a matter of being able to state it and apply it, but of understanding its relation to other theorems (such as 1.46, which enables us to construct the squares needed in the diagram that 1.47 employs in its proof).⁵⁵ Likewise, scientists’ knowledge of the principles of their sciences is not simply a matter of knowing how to state them and knowing that they are indemonstrable, but knowing them as the principles of the proofs in which they are found, that is, knowing which scientific propositions depend on them and how they do. Perhaps this is what is meant by ‘the entire universal’ (100a6). Aristotle’s scientists have knowledge; what they say is true is true; in fact, it is the truth, the whole truth and nothing but the truth. Consequently, they all agree. They all have the same rich knowledge of the concepts involved. In this context, given the purpose of the Posterior Analytics, problem (2) simply does not arise. Aristotle is describing expert knowledge; the fact that non-experts disagree among themselves and with the experts is simply irrelevant.

Problem (3) receives a partial solution from the same consideration. Of course experts agree – what they know is the same. But this answer still leaves open the question how the opinions of scientists, based as they are on epagōgē, are guaranteed to be correct. Does Aristotle give us any reason to believe that epagōgē always results in the truth? The answer is that he does not, that this is a gap in Aristotle’s account. I have discussed this matter elsewhere⁵⁶ and will present a brief summary in the following.

The product of epagōgē is not nous: seeing Socrates as a human being or seeing human beings as animals is a long way from having knowledge of any principle of biology. Instead, epagōgē produces noēsis, a term he uses for the apprehension of ‘things that have an explanation different from themselves’⁵⁷ which are obviously not principles. Since nous and epistēmē are said to be the only two infallible cognitive states (An. post. 100b7–8), there is reason to believe that noēsis is fallible. In fact, Aristotle’s explanation of burning lenses (An. post. 88a14–16) is incorrect as is Anaxagoras’ account of thunder as the noise of fire in clouds being quenched (An. post. 94a35).⁵⁸ These false thoughts will have occurred to their authors through epagōgē in the same way that the true thought that fire melts ice occurs to people, but they have been expelled from science because of their incompatibility with other claims which are considered by experts to be true.

I address problem (4) as follows. It is reasonable to suppose that the way we come to recognize Socrates, Callias, and other human beings as similar and consequently apply a common designation to refer to them and others who resemble them is the same as or similar to the way we come to recognize that this dog and that several other dogs are similar in that they have four legs and then conceive that all dogs have four legs. On this reasonable view, all universal facts treated by a science are learned in this way. This is why when Aristotle says that he will talk about how we come to have nous, that is, how we come to know scientific principles as scientific principles (which would be classified as B-universals), he gives as an example the (simpler) example of moving from Socrates to human being to animal, leaving it to us to transfer the point of the example from A-universals to B-universals, from [Socrates, Callias and other human beings are mortal] to [human beings are mortal] to [animals are mortal]. Let us grant him that this is the way we come to have knowledge of scientific facts.

Problem (5) is not so easily dismissed since Aristotle has said no more than that science deals with both kinds of universals (A-universals and B-universals), that universals of both kinds are grasped through epagōgē and that nous is a different kind of grasp of its objects than epistēmē is. The question then is how to decide which of the A-universals and B-universals are basic and which ones are not. The basic A-universals will be the basic subjects of the science whose existence is posited, and the basic B-universals will be the basic facts about those subjects that constitute their essence and that are stated in the definitions of the subjects. This basic information is given in the existence-claims and definitions that are two kinds of Aristotelian scientific principles.

We⁵⁹ can imagine someone going out and first collecting all the facts⁶⁰ of a science (in practice, this would be difficult or impossible to do). Some will be principles and the rest will be provables. We find out which ones are which in part by relying on the nature of demonstrations, which are syllogistic proofs that show how the provables follow from the principles. The first part of this task is to determine which facts can be deduced from which. This is the work of dialectic, something that is not mentioned in the Posterior Analytics but that is clear enough from the Topics. The key claim there is that ‘this task (namely, discussing the principles of sciences) belongs properly, or most appropriately, to dialectic; for since dialectic is the art of examination (exetastikē) it leads the way to the principles of all disciplines’ (Top. 101b2–4). The examination will include the work of determining which facts follow from which. So far there is no mystery, just the systematic application of the rules of logic with a critical eye. At this point some (perhaps the majority) of universal propositions have been eliminated from consideration as candidates for principles.

But there is more to it! In order to reach the principles, we also need to determine which facts are more intelligible than others because, as Aristotle notes, there are cases where two facts are deductive consequences of one another. More precisely there are cases where two syllogisms are such that the conclusion of each is a premise of the other.

Aristotle discusses this kind of case in Posterior Analytics 1.13. He gives an example: some heavenly bodies twinkle, but a few do not. The non-twinklers are unusual in another way too: they are planets, identifiable as such by their changing positions in the sky relative to the twinklers (the stars). Also, the planets are nearer the earth than the stars are. This was known to Aristotle. Here are two facts, then, in astronomy:

The planets do not twinkle.

The planets are near the earth.

In fact, says Aristotle, all non-twinkling heavenly bodies are near the earth and all heavenly bodies near the earth do not twinkle. On this basis we can form two parallel arguments:

A	(A1) All heavenly bodies near the earth do not twinkle
	(A2) The planets are heavenly bodies near the earth
	(A3) So, the planets are heavenly bodies that do not twinkle
B	(B1) All heavenly bodies that do not twinkle are near the earth
	(B2) The planets are heavenly bodies that do not twinkle
	(B3) So, the planets are heavenly bodies that are near the earth

Both arguments have (let us suppose) true premises and true conclusions, and in both the reasoning is valid. But the conclusion of each argument is a premise of the other. It follows that arguments A and B cannot both be demonstrations. (If argument A is a demonstration, then A2 cannot be demonstrated, and so argument B, whose conclusion is identical with A2, cannot be a demonstration. And vice versa for argument B.) In our search for the principles of astronomy we need to decide whether A2 is a principle for the proof of B2 or vice versa. But how can we decide? Neither logic nor dialectic can help. Aristotle says that A is a demonstration and B is not. Why? Because demonstrations do not just establish that the conclusion is true, they explain why it is true. Argument A tells us that the reason why the planets do not twinkle is that they are near: it is because they are near that they don’t twinkle. Argument B tells us that the reason why the planets are near is that they don’t twinkle: they are near because they don’t twinkle. Aristotle thinks that it is obvious that this is false. No reasonable person would suppose that it is the fact that they do not twinkle that causes the planets to be near!

On the other hand, that they do not twinkle is more obvious to us than that they are near, and we might well say that we can tell that they are near because we see that they do not twinkle. This is more familiar to us. However, Aristotle holds, in general the facts used in demonstrating these obvious facts are not so easily observed, but they are more basic and it is because they are ‘better known in nature,’ that is, more intelligible, that they explain the obvious facts which are ‘better known to us,’ that is, more familiar.

The primary text for this distinction between ways of being ‘better known’ (gnōrimōteros) is An. post. 71b33–72a5:

Things may be prior and better known in two ways: for what is prior in nature and what is prior in relation to us are not the same, nor what is better known and better known to us. I call prior and better known in relation to us the things that are nearer to perception, and prior and better known without qualification the things that are farther. The things that are farthest [from perception] are those that are most universal and those that are nearest are the particulars; and these are opposed to one another.

Since scientific principles are better known ‘without qualification’ (haplōs), or ‘better known in nature’ (tēi phusei) – two designations that Aristotle uses, this passage might leave us wondering whether it is possible for ‘us’ to know what is ‘better known in nature’ and leaves it an open question how we can do so. The situation is clarified somewhat by the statement in the Metaphysics that learning proceeds through that which is less well known in nature to that which is better known (sc. in nature) and that we should strive to make what is better known in nature better known to ourselves (Metaph. Z.3 1029b1–8). What is ‘better known to us’ refers not to a condition we cannot surpass but to our initial condition where we are perceiving things (acquiring knowledge that something is the case) and do not yet have explanations (knowledge of why it is the case), in particular the explanations expressed in scientific demonstrations. As we learn we advance from the former condition to the latter.⁶¹

This means that by ‘better known to us’ he means what is more obvious to non-experts. When we have become experts, we ‘see’ things differently. What is better known in nature has become better known to us. The things that were originally better known (that is, better known to us) are now understood to be less well known, not because they are less obviously true, but because we now know them better since see that they are no longer isolated facts to be treated individually, but are explained by other, initially less obvious facts, and so are incorporated into a systematically organized body of knowledge. As Aristotle says elsewhere, ‘a person who has knowledge through demonstration must know the principles better and be more convinced of them than of what is being proved’ (An. post. 1.2 72a37–39). In effect, in the terms of Aristotle’s distinction between these two ways of being better known, the ultimate explanation is what is best known in nature.

For our purposes this represents a solution to our final problem, how epagōgē⁶² produces the universal in us. Experts see things differently than ordinary people do. What is better known to them is not the same as what is better known to ordinary people. Epagōgē leads us to see individuals as individuals of a kind (to see Socrates as a human being), but also to see some kinds as species of other kinds (to see human beings as animals) and also to see humans as mortal, and it is also what enables experts to see some facts as related in various ways to others (to see that Socrates is mortal because he is a human being, to see this loss of light from the moon as occurring because the earth is blocking the sun’s light), and, ultimately, to see some facts as basic to an entire science.

The result so far is that scientific principles are what the experts agree are most intelligible. Aristotle never gives a satisfactory account of the criteria of intelligibility, saying only that things nearer to perception are prior and better known in relation to us, while things prior and better known in nature are things that are further from perception and more universal, as opposed to being nearer to perception and particulars.⁶³ That these criteria are inadequate is shown by the ambiguity of the expression ‘better known to us.’ As noted above, ‘us’ refers to laypersons as opposed to experts, but how can anyone tell whether the person to whom something is better known is a layperson or an expert? Certainly not when there is disagreement. In many cases a non-expert will not be able to decide which of the competing claims is true, let alone which is a basic truth. Universality too fails as a criterion. One need not be a biologist to know the universal [animal] to some degree – and who is to determine what degree of knowledge or knowledge of which facts is needed for expertise?

The result of this discussion is that if there are experts they have the requisite knowledge – they have the full knowledge of all the ‘propositions’ in the science – which ones are principles, which ones are provables, and all their interrelations – while the rest of us do not, even those of us who think that they do. This is also true in the hypothetical case where different people or groups of people agree on all the facts but identify differing sets of facts as principles. Aristotle wisely does not try to adjudicate such cases, and this accords with the answer given above to questions (1) and (2). The Posterior Analytics does not tell us except in the most general terms how to reach scientific principles; it simply sketches how achieved sciences (were there ever to be any) would be structured. However, the discussion of the present section accounts for the ideal of expert knowledge and the conditions of its realization. Moreover, since experts agree in ‘seeing’ that the principles are fundamental and since having full knowledge of a science involves knowing all the facts and all the demonstrations, and experts have this full knowledge, they have what can be considered full knowledge of the concepts involved.

5 Back to Concepts

The account just offered of how one might attain knowledge of scientific principles suggests the following view of concepts. Different people have different concepts of the same thing. In that the earliest concept I had of human beings stemmed from different acts of perception than you had, my earliest concept of human beings was different from yours. As we grew up our concepts became more detailed, and they came to overlap considerably but (it can be argued) not completely. In particular, our concepts enable us to explain (rightly or wrongly) some features of human beings as due to others and also enable us to anticipate and predict certain behaviours and events. Also, our concept of human beings includes that we (as human beings) have characteristics that (some) other animals have too, and our concept of human being includes that human beings are animals. So far, this all occurs without any pretension to knowing everything about human beings, but it does prompt the thought that knowledge of human beings requires (or perhaps is just part of) knowledge of animals. This can be seen as our having successively different concepts of human beings. In this sense different persons have different concepts of the same thing or fact, and a person’s concepts can change, for the better or for the worse.

If we study biology, we learn more about living things in general, and much of what we learn gives us further knowledge of human beings. Our concept of human being continues to change and develop towards a scientific understanding. If biology were complete in the same way that one can have complete knowledge of simple games like tic-tac-toe, it could be fully mastered – a person who mastered it would know all the relevant facts and all their relations (which ones depend on others and how the explanations go). At this stage, one would have full knowledge of the concept of human being and this corresponds to expert knowledge of a complete Aristotelian demonstrative science.⁶⁴ In this sense, Aristotle’s discussion of epistēmē and related issues in the Posterior Analytics and elsewhere can be seen as pointing the way to knowledge conditions for the concept of something as obtained through our own private concepts.

6 Conclusion

Aristotle’s discussions in Posterior Analytics 2.19 and Metaphysics A.1 do not explicitly address the topic of concept acquisition and it is not clear that concepts played a significant role in his thinking about epistēmē and nous. However, his two accounts of the progress from sensation to memory to empeiria and beyond, which necessarily involve progress from having cognition of particulars to having cognition of universals, can be employed to disclose what his views might have been about how we acquire concepts, and they have three interesting and suggestive implications for how we think about concepts: First, the idea of growing concepts – that our concept of something can change and in favourable cases can become more detailed and more explanatory; second, that there are concepts with propositional content; third, that there are concepts of individuals as well as of universals. These are consequences of identifying Aristotle’s noēmata (and ennoēmata) with concepts. On the other hand, if we hold that there are concepts only of what Aristotle would call universals, then Aristotle’s view that there are noēmata of particulars might be taken to indicate that the attempt to discuss concepts in connection with Aristotle’s logical works is mistaken.