1. Introduction
More than any other philosopher, Hume attempted to analyse our conceptual, propositionally structured thought solely in terms of our ultimate awareness of nothing but objects, whether sensory impressions or their imagistic copies, “ideas.” Because they are so fundamental to his Treatise of Human Nature and to philosophy, Hume’s Copy Theory of impressions and ideas and his account of abstract ideas have commanded considerable critical attention. It may appear foolish to venture a further word on the topic, more so the word proffered here: that everything required to recognize the significant limits of the Copy Theory and of Concept Empiricism is provided by Hume himself in his account of abstract ideas. Hume’s account of the generality of thought includes determinables; however, it does so by appeal to further resources solely of the imagination. This issue is not exhausted because neither Hume’s critics nor his expositors have posed it quite rightly. To establish these points requires strictly internal critique, for reasons which become apparent as we proceed.
To Locke’s empiricism, Leibniz famously retorted: “Nihil est in intellectu quod non fuerit in sensu, excipe: nisi ipse intellectus.”Footnote 1 Whatever critical implications this claim may have for Locke’s Essay, it has none against Hume’s Treatise, because his Copy Theory of sensory impressions and ideas is paired with his three laws of psychological association, which Hume attributes to human nature. Hume’s laws of association do not result from sensory impressions, not even in combination with their ideational copies. Instead, these laws indicate how we process sensory impressions and ideas; they form the core of Hume’s account of our “intellect,” so far as this may be contrasted (briefly, for the sake of discussion) with our “senses” (per Leibniz). The key questions concern precisely how these two components of Hume’s view are (or are best) conjoined, and to what extent their proper conjunction provides an adequate empiricist account of the generality of thought. However these matters may ultimately be understood, sensory impressions and their image-copies (ideas) are at most occasioning causes of abstract or general ideas; they are not and cannot be their sufficient causes.
Critics have long argued that forming a relevant Humean association amongst resembling impressions and ideas requires already having the concept which provides the proper classification by which to include only relevant impressions or ideas, to the exclusion of others.Footnote 2 To this ‘Standard Objection’ (as I shall call it) Garrett rightly replies:
This objection, however, misses the point of his naturalistic theory of concept acquisition, which requires only that similar instances be able to produce similar effects in the mind, including the idea of a similar word. For Hume, similar causes can often produce similar effects, whether outside the mind or inside it, prior to the existence of any concept of the respect of similarity.
Humean abstract ideas thus differ from ideas of particulars not through their intrinsic character but through their causal and functional relations.Footnote 3
A further point should be noted: the Standard Objection has the untoward consequence that all concepts are a priori, and that no concept can be acquired or learned through experience, because ex hypothesi forming or noticing the relevant relations—e.g., specific resemblances in kind or characteristic(s) among sensed particulars—requires already possessing and correctly using the concept in question. For any concept to be learned or acquired on the basis of experience, as Garrett notes, similar sensed or recollected instances must produce similar (preconceptual, implicit) effects in the mind, which provide a necessary basis for forging the relevant concept (classification) and designation (name). However, neither the Standard Objection nor these rebuttals settle the central issue, because they do not engage it.
The central issue is whether these “effects” can be accounted for in accord with Hume’s Copy Theory, Concept Empiricism and three official laws of psychological association, or does our imagination generate these effects regardless of those official tenets of his empiricism? This issue turns, in part, on the contrast between sensory or perceptual specificity and classificatory specificity, and in the latter case, between determinate and determinable classificatory specificity. This tripartite distinction of forms of specificity, it shall be argued, is required for an adequate account of our human form of mindedness, and of Hume’s descriptions of it. There are no simple particulars whatever, in the sense of a particular having only one quality or characteristic. This holds for physical particulars, for sensory impressions and for sensory images, such as Hume’s sensory ideas qua copies of impressions. Hume rightly stresses this point in his criticism of Locke’s account of abstract ideas. Particulars are determinate because each instantiates a specific plurality of determinate characteristics and spatial or temporal parameters (or both).Footnote 4 In contrast, classificatory specificity is a function of often fine-grained discrimination, and our capacity to specify fine-grained classifications, each of which picks out only one or a very few classificatory contrasts, whilst also disregarding irrelevant differences amongst their instantiations by particulars. Such classifications may be determinate, so that their only instances are exact replicas of, e.g., some one specific shade (hue, saturation and brightness) of colour, such as “dusty periwinkle blue,” as defined by a specific colour manual.Footnote 5 However, classificatory specificity also includes determinables, classifications which admit of various ways or degrees to which they may be instantiated, which can be very selective of what characterises relevant instances. Determinable classifications include, e.g., “coloured” or “blue” (in contrast, inter alia, to dusty periwinkle blue). I shall argue that Hume’s Copy Theory and Concept Empiricism are limited to specifying determinate classifications, whereas his account of the imagination’s capacity to associate particular perceptions by various “resemblances” is not thus limited, but includes also determinable classifications. This is the key to Hume’s account of abstract or “general” ideas, of general terms and of the generality of thought. However, Hume provides no explanation of this capacity of imaginative association, and so provides no distinctively empiricist explanation or account of it.
Ultimately, on Hume’s account, this imaginative capacity to discriminate and to identify determinable classifications properly belongs to the understanding. Hume specifies the imagination’s associative capacity regarding resembling characteristics by ascribing to the imagination all the comparative and discriminatory powers ascribed by his predecessors to the understanding (or to the “intellect,” as Leibniz says). Identifying “resemblances” regarding either perceptually specific or classificatorily specific shared traits, either determinate or determinable, is the most general and established regularity of the imagination Hume identifies, and it is fundamental to his entire philosophy. Hume’s definition of the understanding, as “the general and more establish’d properties of the imagination” (T 1.4.7.7, SBN 267), entails that this remarkable, discriminatory, classificatorily specific power of imaginative association via fine-grained and also determinable resemblances ultimately belongs to human understanding. Because Hume specifies the imagination’s associative and discriminatory power in terms of classificatorily specific determinables, the “revival sets” of associated idea-instances are strictly irrelevant to those classifications and to the generality of terms, as Hume himself acknowledges. The relevance of any revival set or of its members is defined by the classificatory specificity of the imagination’s associative (and discriminatory) “resemblance” (classification), not vice versa, and the imagination’s discriminatory, classificatory power suffices unto itself for the generality of thought. Relevant examples we may recall ad libitum, though for thinking they are largely dispensable. (They become most relevant if our grasp of a term, concept or particular event or example falters.)
In brief, Hume’s account of the generality of thought and of language all depends upon ascribing to the human power of imaginative association the classificatorily specific, discriminatory capacity to pick out only one relevant characteristic (or a very few relevant characteristics)—some determinate and others determinable—under the casual, ambiguous and misleading headings of “resemblance,” “psychological association” and “custom.” What needs to be explained is how these psychological associations include often very fine-grained, discriminatory though determinable classifications. For such an explanation, Hume’s Copy Theory, Concept Empiricism and three psychological laws of association, whether singly or combined, do not suffice. Concept Empiricism is the thesis that every (legitimate, genuine, significant) concept is either a logical term, a name for a simple perceptual quality, or can be defined solely and exhaustively by combinations of these two kinds of terms. Twentieth Century empiricists had to relearn how severely limited is this thesis about conceptual content (or mutatis mutandis linguistic meaning), after which it vanished from the tenets of recent empiricism. In practice, and very clearly in theory too Hume etched the limits of Concept Empiricism, though often without expressly noting his achievement in this regard. This occurs plainly in “Of Scepticism with regard to the senses,” where Hume must introduce several propensities to form specific beliefs which are key components of our belief in “body” (i.e., physical objects). Each of these beliefs contains specific conceptual content which cannot be defined in accord with Concept Empiricism; this is why Hume must augment his three official laws of psychological association with three further propensities.Footnote 6 The conceptual innovations each of these propensities involves is obscured, though not eliminated, by speaking broadly (as Hume does) of “belief.” However, the concepts required to specify the content of these beliefs cannot be defined in accord with Concept Empiricism; their failure to fulfill the requirements of Concept Empiricism is Hume’s ground for denouncing them as “fictions.” Yet these concepts, illegitimate as they officially are according to Concept Empiricism, are ineliminable components of the generality of human thought and of the conceptual possibility of pervasive human beliefs. Part of Hume’s brilliance is to demonstrate these results, despite their departure from his official empiricist principles.
Hume’s account of our ideas of space and time have long been regarded as anomalies in, if not exceptions to, Hume’s theory of ideas and his account of the generality of thought. I argue that these ideas are not anomalous, but rather are typical of Hume’s account of the generality of thought in terms of “general ideas,” which account ultimately undermines Hume’s official empiricist commitment to the sufficiency of the Copy Theory, Concept Empiricism and the three official laws of psychological association.
To justify these results, I first disambiguate (to the extent required) Hume’s term “idea” (§2). I then reexamine Hume’s “idea of existence,” to show some key equivocations between “ideas” as objects (images, particular perceptions) and “ideas” as classifications which are crucial to Hume’s account of the generality of thought (§3). The fundamental role played by these equivocations in Hume’s account are then exposed by reexamining Hume’s account of abstract ideas (§4), his inclusion of determinables within his account of “general ideas” (§5), and his accounts of distinctions of reason (§6), of the idea of equality (§7) and of the ideas of space and time (§8). I corroborate these findings by considering a potential counterexample (§9). From these considerations, I conclude that Hume’s account of the generality of thought is fundamentally rooted in judgmental discriminations of determinable kinds which cannot be explained or otherwise accounted for by appeal only to particular perceptions of any sorts, whether singly or in combination (on the basis of Hume’s three laws of psychological association). Hume is, thus, not only the great Modern exponent of the Copy Theory and of Concept Empiricism, he is also one of their most instructive critics (§10).Footnote 7
2. Disambiguating Hume’s “Idea”
Descartes’s, Locke’s and Hume’s accounts of sensory ideas were designed to fill a crucial role as sensory foundations of human knowledge. To fill this role, each of these philosophers (along with many contemporaneous logicians) assimilated the awareness of some (mental) object to the awareness that the object in question has certain manifest characteristics. Their common thesis, that sensory ideas are exactly what they seem to be, assimilates the characteristics an idea exhibits to characteristics one believes (or perhaps knows) that idea to have. Put otherwise, they conflated objectual and propositional senses or kinds of “awareness.” This conflation commits them, in effect, to aconceptual “knowledge by acquaintance” of sensed particulars, whether these be physical objects or sensory impressions (in Hume’s sense too). To note this conflation is not, as such, an objection. Neither Descartes nor Locke nor Berkeley attempted to account for all human thought in terms of our awareness of (literal, if mental) objects. Hume, however, worked out the presuppositions and implications of this kind of theory in detail. As an arch empiricist, Hume disregarded—and so implicitly rejected—Berkeley’s distinction between “ideas” and “notions,” such as the notion of spirit, for which no sensory instances can be given.Footnote 8
The distinctive character of Humean ideas can be framed by using Descartes’s distinction between the “formal” reality of ideas and their “objective” reality. According to Descartes, the “formal” reality of an idea concerns the very being of the idea itself, as an extant mode of a thinking substance which is someone’s mind. The “objective” reality of an idea concerns the representational capacity of an idea, its capacity to represent something, typically something distinct from itself, and to represent it as having some characteristic(s) or other. Central to Descartes’s rationalism is his focus upon the “objective reality” or representational capacity of ideas to account for propositional modes of thoughtful, conceptually structured awareness of things (very broadly speaking) and their features. As if commenting on Descartes’s distinction between the formal and objective reality of ideas, Hume states:
… to form the idea of an object, and to form an idea simply, is the same thing; the reference of the idea to an object being an extraneous denomination, of which in itself it bears no mark or character. (T 1.1.7.6, SBN 19)
As a radical empiricist, Hume sought to explain any supposed “objective reality” of ideas solely in terms of the formal reality of ideas as extant mental occurrences, as literal objects of awareness (albeit functioning within a suitable associational context). This is explicit in Hume’s Copy Theory and in Concept Empiricism.Footnote 9 Hume expressly (though not univocally) equates “thoughts” with “ideas” in his specific sense of the term, according to which they are particular sensed objects (“perceptions”; T 1.1.1.1–3, SBN 1–2). I shall refer to this as Hume’s objectual account of ideas. To keep Hume’s official doctrine in focus, and to avoid confusing what Hume may have conflated, I shall use the term ‘idea-object’ to designate Hume’s “ideas” as literal objects of awareness, as he officially conceives them.Footnote 10
I shall argue, to the contrary, that Hume’s account of distinctions of reason affords an account of the generality of thought only because Hume simply builds into the power of human imagination the capacity to associate particular impressions or idea-objects by classificatorily specific yet determinable “resemblances.” The relevant “associations” are determinable classifications; to their use in thought or discourse, associated idea-objects are irrelevant, because the relevance of any such idea-objects is specified by the classification, not vice versa. We can begin to appreciate this point by reconsidering Hume’s “idea of existence.”
3. Hume’s Idea of Existence
Central to Hume’s objectual account of ideas (‘idea-objects’) is this thesis: “What consists of parts is distinguishable into them, and what is distinguishable is separable” (T 1.2.1.3, SBN 27), and conversely, that only what is separable is distinguishable (T 1.1.7.3, 1.2.3.10, SBN 18, 36). This is Hume’s “Separability Principle.” Officially, any distinction we can make in thought depends upon our identifying two different idea-objects, each of which instantiates only one of the two distinct ideas. This tenet requires Hume to insist, as he does, that the idea of existence or of being is identical with any idea whatever that we form or recall (T 1.2.6.4, SBN 66–7).Footnote 11 This involves a philosophical sleight of hand, signalled by Hume’s revealing statement that:
The idea of existence … is the very same with the idea of what we conceive to be existent. (T 1.2.6.3, SBN 66)
Two important points are suggested by this formulation. First, Hume’s use of the verb “conceive” is crucial. Officially, according to Hume, to “conceive” something is presently to have (within oneself qua bundle of impressions and idea-objects) a something-idea, a determinate idea-object. Literally, to conceive of something is to think of it in some definite way. To what extent does the very “idea of existence” involve an act of conceiving, or (mirabile dictu) a concept of existence? Second, Hume’s formulation uses the grammatically ambiguous term “of,” which can mark either a subjective or an objective genitive. An objective genitive would, in this context, underscore the intentionality of ideas, that ideas represent something other than themselves, and represent it as having certain characteristics. A subjective genitive would, in this context, underscore the manifest (sensory) characteristic(s) of the very idea-object in question. To keep Hume’s official view clearly in focus, namely, his repudiation (in effect) of the “objective reality” or intrinsic representational capacity of ideas noted above (§2), he would have done better to use ‘that’ rather than “of,” viz.:
»The idea of existence, then, is the very same with the idea that we conceive to be existent.«Footnote 12
Though the questionable term “conceive” remains, this reformulation underscores how seriously Hume must take his identity claim (“is the very same with”) in order to maintain his objectualist account of the idea of existence. Because Hume equates distinguishability and separability, if this identity were correct, the phrase “idea of existence” would be strictly redundant, a pleonasm. If the terms “idea” and “existence” are not redundant, if they are distinct significations, then Hume is not entitled to his identity claim.Footnote 13 Hume recognises this point elsewhere, in another connection (T 1.2.4.27, SBN 50). Nowhere in his discussion of the idea of existence does Hume consider that these two terms are simply redundant.Footnote 14 This suggests his awareness, if perhaps tacit, that the phrases “idea” and “idea of existence” have distinct connotations, even if (officially) they have the same denotation. This phenomenon is central to distinctions of reason, and to Hume’s account of such distinctions (see below, §4).
Notice that Hume’s identity claim is untenable for a very simple reason: each and every present (occurrent) idea-object exists, no more and no less than any and every other idea-object (whenever it occurs). Though each idea-object exists, the idea of existence cannot be identified—that is, numerically equated—with any one present idea-object, because it is equally the case that any and all other idea-objects exist whenever they occur. Neither can the idea of existence be identified with any plurality of idea-objects, because each occurrent idea-object exists regardless of whether any others do. In brief, the idea of existence cannot be defined purely extensionally. To define the idea of existence requires, not numerical identity, but qualitative or specific identity (i.e., identity of kind). Elsewhere Hume recognises this distinction, stating that:
… we commonly [are] able to distinguish pretty exactly betwixt numerical and specific identity, yet it sometimes happens that we confound them, and in our thinking and reasoning employ the one for the other. (T 1.4.6.13, cf. 1.1.5.10, SBN 257, cf. 15)
Hume’s examples of such conflation (“confound[ing]”) concern our common assumption of the numerical identity of perceived objects, despite gaps in our (putative) experience of them (T 1.4.6.13, SBN 257). He does not note his own confounding of numerical and specific identity in his account of our idea of existence. More importantly, he cannot recognise this without recognising the inadequacy of his objectualist account of thought. Any adequate account of human thought must account for our recognition, our understanding, of specific identities and of specific differences, i.e., what is identical or different in “species” or kind. Such identities and differences are based upon specific features or characteristics of particulars. The initial problem for an objectual account of thought is that, like any object, regardless of how “simple” it may allegedly be, any idea-object either is, or has, a complex of features, characteristics, or (as I shall say) aspects. A simple visual idea-object features shape or figure, colour (which can be more precisely distinguished into the three aspects of hue, saturation and brightness), size, duration and location within one’s visual field. A simple gustatory idea-object features a specific flavour (which is typically a complex of flavours), intensity, and duration; and likewise in other cases.Footnote 15 This is true of any extant idea-object and whatever characteristics it displays, whether a colour or a shape or an auditory tone, etc. Hume himself insists upon the complex determinacy of idea-objects when he rejects Locke’s notion of abstract ideas (T 1.1.7.2–6, SBN 17–19). Identifying or recognising aspects of any one idea-object requires identifying or recognising specific characteristics of and differences among its aspects. Noting the specific character of such aspects involves identifying them by individuating them within the one determinate idea-object which exhibits them. In Modern terminology, such distinctions are distinctions of reason, distinctions that can be made only in thought, because there is no real distinction (into separable parts or components) within the object thought about, or so conceived. Hume, of course, recognises the limits of his Separability Principle in his account of distinctions of reason, which is central to his account of abstract ideas. Examining Hume’s account of abstract ideas provides the proper context for considering the bearing of distinctions of reason upon minima sensibilia, which might be suggested as exceptions to the otherwise complex determinacy of idea-objects.
4. Hume’s Account of Abstract Ideas
Hume rejects Locke’s account of “abstract ideas” because, like the impressions of which they are copies, ideas can only be particulars and so are determinate in each of their various respects. Hume insists explicitly that whatever exists is manifoldly determinate, ideas no less so than impressions or any other kind of particular object (if we could but conceive or perceive any other kind of object) (T 1.1.7.2–6, SBN 17–19).Footnote 16 Since impressions and idea-objects literally are objects of thought and awareness, any idea-objects we have can only be manifoldly determinate; they are not determinables. Because Locke’s “abstract ideas” would have to be indeterminate images (or vague though very image-like), Hume rightly rejects them. On Hume’s account, an “abstract idea” is simply a manifoldly determinate idea, which is used generally, that is, indifferently, to stand for—literally, to recall—any or all idea-objects of that kind or, as he says, “species.” In this, Hume follows Berkeley (part way) and adopts radical nominalism (T 1.1.7.1, SBN 17).
The key points of Hume’s account of abstract ideas may be quoted verbatim:
Abstract ideas are, therefore, in themselves individual, however they may become general in their representation. The image in the mind is only that of a particular object, though the application of it in our reasoning be the same as if it were universal. (T 1.1.7.6, SBN 19)
This application of ideas, beyond their nature, proceeds from our collecting all their possible degrees of quantity and quality in such an imperfect manner as may serve the purposes of life …. When we have found a resemblance[5] among several objects, that often occur to us, we apply the same name to all of them, whatever differences we may observe in the degrees of their quantity and quality, and whatever other differences may appear among them. After we have acquired a custom of this kind, the hearing of that name revives the idea of one of these objects, and makes the imagination conceive it with all its particular circumstances and proportions. But as the same word is supposed to have been frequently applied to other individuals, that are different in many respects from that idea, which is immediately present to the mind; the word not being able to revive the idea of all these individuals, only touches the soul, if I may be allowed so to speak, and revives that custom, which we have acquired by surveying them. They are not really and in fact present to the mind, but only in power; nor do we draw them all out distinctly in the imagination, but keep ourselves in a readiness to survey any of them, as we may be prompted by a present design or necessity. The word raises up an individual idea, along with a certain custom, and that custom produces any other individual one, for which we may have occasion. But as the production of all the ideas, to which the name may be applied, is in most cases impossible, we abridge that work by a more partial consideration, and find but few inconveniences to arise in our reasoning from that abridgment. (T 1.1.7.7, SBN 20Footnote 17)
[I]t is after this manner we account for the foregoing paradox, that some ideas are particular in their nature, but general in their representation. A particular idea becomes general by being annexed to a general term; that is, to a term which, from a customary conjunction, has a relation to many other particular ideas, and readily recalls them in the imagination. (T 1.1.7.10, SBN 22)
This Berkeleian account Hume proclaims to be “one of the greatest and most valuable discoveries that has been made of late years in the republic of letters” (T 1.1.7.1, SBN 17). The core of Hume’s account is that resembling ideas are associated by custom, which allows us to name them with a word. In turn, that “word raises up an individual idea, along with a certain custom, and that custom produces any other individual one, for which we may have occasion” (T 1.1.7.7, SBN 20). Hence no idea-object as such has generality, only its use and its associated term do, due to the customary association of any one idea-object with a range of resembling idea-objects and occurrent impressions (if any), where these customary associations result from Hume’s three laws of psychological association.
Hume’s linguistic account of the general use of particular idea-objects appears to be in tension with his objectualist account of thought. Hume’s objectualist account of thought cannot account for our awareness of relations, because relations among particulars (whether egocentric or otherwise) are not themselves particular objects, and so cannot produce Humean idea-objects, certainly not in accord with the Copy Theory. (This claim will be further corroborated below.) It appears that Hume doesn’t regard the introduction of general terms as compromising his objectualist account of thought because he regards general terms simply as linguistic tags for groups of associated idea-objects. There are, however, cases in which Hume does attempt to treat “ideas” (idea-objects) as general, e.g., the ideas of existence, space and time.
I do not criticise Hume’s nominalism, but instead argue that our recognition of and thought about relations is an intellectual achievement effected by the power of human imagination to discriminate, to identify and to classify relevant, classificatorily specific, determinable and often very fine-grained “resemblances” among particular perceptions (whether impressions or idea-objects). Such discriminatory classification is required for any propositional awareness that some particulars are related in some specific way(s), where this propositional awareness cannot be reduced to our objectual awareness of particular objects, whether singly or distributively.Footnote 18
Officially, one idea naturally introduces another by a “gentle force” of association, whether by resemblance, contiguity or cause and effect (T 1.1.4.1, SBN 10). Many readers have followed Hume’s Newtonian analogy (T 1.1.4.6, SBN 12–3) and have taken “force” literally here, as some kind of causal attraction among idea-objects. However, given his scepticism about causal powers (however interpretedFootnote 19), Hume cannot admit such a force into his account of human nature. Within Hume’s own account of human nature, this gentle force can only be the figurative force of psychological association, for which Hume can only account in terms of constant conjunctions, including those he considers cause-effect relations (T Appendix ¶20, SBN 633).Footnote 20
This accords with the many actions Hume ascribes to us in our developing any customary association suitable for a significant name. In the passages just quoted, Hume speaks of our “application” of an idea-object in “our reasonings … as if it were universal,” of our “collecting all possible degrees of quantity and quality” of idea-objects; of our having “found” resemblances; of our “acquiring” a custom by “surveying” a range of resembling idea-objects; of our “apply[ing]” a common name to a group of resembling “objects”; of our “observ[ing]” differences in degree and quality among those objects; of our “keep[ing] ourselves in a readiness to survey any of them, as we may be prompted by a present design or necessity”; and of our “suppos[ing]” that “the same word [….has] been frequently applied to other individuals,” that is, other resembling individuals of the same specific kind.
These phrases all connote intellectual activities, indeed intellectual achievements; more such phrases are noted below. Hume’s use of such phrases, which so clearly connote mental acts, are central to the Standard Objection to Hume’s account mentioned and rejected above (§1). However, Hume’s phrases are significant in a more fundamental regard: each of them testifies to how focussed, fine-grained, selective or discriminatory can be our attention to very specific features of particulars and to “resemblances” between particulars in any specific regard. Our attention can, and by Hume’s own report often does, exhibit classificatory specificity rather than, or in addition to, perceptual specificity (the manifold determinacy of any particular of which we can be aware; on Hume’s view: impressions or idea-objects). This our attention can do, on Hume’s account, only because our imagination has the capacity to identify very selective, fine-grained “resemblances” among particulars, where these “resemblances” involve classificatory rather than perceptual specificity. The perceptual specificity of idea-objects only affords classifications of exact replicas; this is a direct consequence of their manifold determinacy, upon which Hume rightly insists. This is an important form of classification, but it cannot afford classifications of any one feature of idea-objects. Hume allows for “secondary ideas,” which are copies of idea-objects, which as such are themselves idea-objects. A secondary idea-object may copy only a subset of the determinate qualities of its original and originating “primary” idea-object. This is an advance in selectivity, though because secondary idea-objects, too, are particular perceptions (they are particulars we can and, on Hume’s view, must be able to sense), they are equally determinate with respect to their subset of determinate qualities as their originating primary idea-objects. Consequently, their gain in selectivity is, in principle, no gain in determinability; secondary idea-objects, too, can only classify exact (if partial) replicas.
Can secondary, or if need be, tertiary or quaternary idea-objects have only one characteristic, and so serve to classify only one characteristic? Answering this question on Hume’s behalf requires preserving his core empiricist view that, like impressions, idea-objects are objects (however vivid or faint) of sensory, or we may allow, quasi-sensory awareness: they must appear to us to bear specific, determinate sensory qualities, for they are all “perceptions.” If this requirement were relaxed, Hume’s objection to Locke, the distinctiveness of his account of abstract ideas and his objectual account of thought would all be abandoned. Abstract ideas, too, are perceptions within a bundle which is some one of us, and so are perceptible (or rather, “sensible”); to rescind this premise is to rescind Hume’s Copy Principle. Any answer to the above question will be controversial, but I will venture an answer because however this question may be answered, it serves to underscore why Hume’s Copy Theory and his Concept Empiricism cannot suffice for his account of abstract ideas or the general use of terms.
The most plausible candidates within Hume’s framework for tertiary (or n-ary, if one wishes) idea-objects which instantiate and so serve to classify only one specific characteristic are these: a minimum visibilium of some one shade of colour, or a minimum tactilium of some one texture, or a minimum auditorium of some one pure tone (say, one generated by a high-quality electronic oscillator, so that overtones, timbre or “sonority” are eliminated or minimized so far as possible).Footnote 21 We begin by omitting considerations of intensity and consider only the qualitative purity of these minima sensibilia, and whether they are indeed “simple” in the sense of having (and so serving to classify) only one characteristic. None of these minima sensibilia are in this sense simple. Any determinate colour exhibits a specific, determinate hue, saturation and brightness. The distinctions between these aspects of any colour cannot be rejected or neglected for the reason that they are not “visually apparent.” Though many people may not distinguish them, they can be significantly distinguished (within some range of colours for most anyone with colour vision) with a modicum of training and attention, that is, by becoming accustomed to differentiating these aspects of any one perceived colour. Differentiating them, of course, requires drawing distinctions of reason. Those who work with colour draw these distinctions habitually.
The other suggested minima sensibilia are no more favourable to the question of mono-qualified n-ary idea-objects within Hume’s Copy Theory. The proposed minimum tactilium will have not only a distinct, determinate texture of, say, some determinate degree of smoothness, roughness or sharpness; it will also have some degree of apparent warmth, coolness or cold. If it is very brief and feels sharp enough to be painful, we may not notice its apparent warmth, but apparent warmth is a tactile quality no less than is texture. Hypothesizing logically possible creatures who somehow have distinct sensory modalities for texture and for apparent warmth abandons the project of a treatise of human nature, and would fail to account for our human form of mindedness, which Hume sought to provide—rightly, I submit.Footnote 22
The best of the three suggestions is the minimum audibilium of some one pure tone. Since by stipulation it is pure, it avoids the problems with the minimum visibilium; since few of us have synaesthesia, it sufficiently minimises the problem with the minimum tactilium. However, a minimum audibilium will have both tone and intensity (loudness), and all minima sensibilia have determinate duration. Hence none of them are qualitatively simple. Because they are particulars, they each have a plurality of distinct, determinate characteristics, which we can distinguish only by making distinctions of reason.Footnote 23
Hume’s rejection of Locke’s notion of abstract ideas requires that his Copy Theory be taken strictly, with the consequence he himself stresses that the characteristics of any and every impression are each determinate, as are likewise the characteristics of any and all of their copies, idea-objects, even if at a secondary or tertiary level of iteration idea-objects come to have ever fewer (and ever fainter?) determinate characteristics. The consequence of this is that, however rich, extensive or finely individuated may be our memorial store of idea-objects, each idea-object is perceptually specific and so exhibits and only thereby “represents,” classifies or identifies its specific plurality of determinate characteristics. Consider someone who might have a memorial store of coloured idea-objects as finely differentiated as the colour chips catalogued by a good manufacturer of decorator interior paints. This multitude of determinate colour-idea-objects could, if conjoined, form an extremely complex colours-idea-object. However, this complex idea-object neither constitutes nor suffices to specify a determinable general idea, “colour.” Likewise, however comprehensive and finely differentiated may be someone’s idea-objects of determinate lengths, conjoining this multitude neither constitutes nor suffices to specify a determinable general idea, “length” (nor (spatial) “extension”). Likewise, however many distinct, determinate temporal sequences of musical notes we may hear and recall, each such determinate-temporal-sequence-idea-object is an idea-object “of” nothing but that determinate duration or sequence. No matter how extensive and finely differentiated may be someone’s (re-)collection of such determinate duration-idea-objects, conjoining them does not suffice to constitute nor to specify the determinable general idea, “temporal duration,” nor “temporal.”
Rather than repeating this point ad indefinitum, note that appealing to Concept Empiricism does not remedy this limit of the Copy Theory. Concept Empiricism defines concepts on the basis of (initially simple) logical combinations of names of simple sensed qualities. (The use of complex logical combinations requires considerable conceptual sophistication, which requires our prior acquisition and competent use of many relevant subclassifications or concepts.) To stay within Hume’s framework, I do not pursue issues about the requirements for naming “simple” sensed qualities, nor any further issues about the simplicity of the requisite qualities. Just as Hume’s view allows the decomposibility and recomposibility of complex idea-objects into, or conversely out of, maximally simple idea-objects, Concept Empiricism allows us to specify and to name various complexes of maximally simple idea-objects, corresponding to no impression we may ever have. Nevertheless, concepts so specified are each of them names of perceptually specific, determinate sensible qualities. For the same reasons given above regarding the Copy Theory, Concept Empiricism can only provide an arbitrarily rich set of names of determinate sensible characteristics, however complexly (re)combined these may be. Yet in this case, too, no matter how rich, comprehensive and finely differentiated may be such named complexes of conjoined determinate characteristics, their names can neither specify nor suffice for constituting their corresponding determinable general term, name or classification. About these limits Hume’s critics—most notably Sellars—are correct.Footnote 24
However, Hume’s expositors have correctly noted that these points do not exhaust the issue of whether Hume can or did account for the generality of thought, though I believe they have not yet properly explicated it. Obviously, I have agreed with Sellars that the key to properly general thought lies in our grasp and proper use of determinable classifications, for which arbitrarily rich and differentiated determinate classifications do not suffice. Yet, what about thoughts (which have surely occurred to the reader) such as “… and any other such-like quality,” that is, “any other resembling characteristic”? Such thoughts involve determinables.
5. Determinability in Hume’s Account of General Ideas
Hume clearly stresses the importance of determinables within the phenomenon of general thought and within his account of “general ideas”:
It is evident, that, in forming most of our general ideas, if not all of them, we abstract from every particular degree of quantity and quality, and that an object ceases not to be of any particular species on account of every small alteration in its extension, duration, and other properties. … though the capacity of the mind be not infinite, yet we can at once form a notion of all possible degrees of quantity and quality, in such a manner at least, as, however imperfect, may serve all the purposes of reflection and conversation. (T 1.1.7.2, SBN 17)
By acknowledging that “we abstract from every particular degree of quantity and quality,” Hume acknowledges that we abstract from any and every determinate degree of quantity or quality in forming “species,” classifications which are distinct from and which include a range of various determinate characteristics, whether of “extension, duration” or “other properties.” Such a “species” classification includes “all possible degrees of quantity and quality” relevant to that species. Two features of Hume’s formulations strongly suggest discriminating and identifying some quantitative or qualitative continuum which embraces any and all “species”-relevant determinate instances, in a way tantamount, if not identical to, a determinable classification. With regard to both quantity and quality, Hume insists that most (if not all) general ideas specify them regardless of their “degree.” A general idea “abstract[s] from every particular degree of quantity and quality,” and indeed it embraces “all possible degrees” of the relevant quantity or quality. These two points indicate a comprehensive form of generality not limited to, and not limited by, any determinate instances of that kind or “species,” no matter how exhaustively enumerated or finely differentiated its determinate instances may be. Such indeterminate scope is characteristic of determinable classifications. Accordingly, Hume’s “general ideas,” his classificatorily specific “resemblances,” include determinables, and so are not limited to determinate characteristics.Footnote 25
How can this be, if Hume unequivocally rejects Locke’s notion of abstract ideas as (in effect) vague and hence impossible idea-objects? How can this be, if the Copy Theory and Concept Empiricism only afford determinate classifications? Whence determinables within Hume’s account of what he here calls “general ideas”? Hume’s answer is disarmingly simple: our imaginative capacity to discriminate and to identify relevant “resemblances” is not limited to determinate resemblances. Our imaginative capacity to discriminate and to identify relevant “resemblances” is not limited to the legitimate products of either the Copy Theory or Concept Empiricism. Our imagination can also discriminate, identify and (allow us to) name determinable resemblances. These determinable resemblances cannot literally be represented (nor instantiated) by idea-objects, because these latter are all determinate (perceptually specific). Hume is, thus, right to introduce the general use of a term which, in effect, designates a determinable classification. I submit it is significant, and not at all accidental, that precisely at the point where he must account for our significant use of determinable classifications within his account of “general ideas” as providing his account of the proper generality of human thought, Hume appeals to language to augment the resources of his Copy Theory and Concept Empiricism.
Hume proposes to link our imagination’s forging of significant general terms to his empiricist Copy Theory and Concept Empiricism by claiming that our significant, general use of any term is parasitic upon our familiarity with and capacity to recall relevant instances, in the form of idea objects. Donald Garrett has conveniently designated these as a general term’s “revival set.” Certainly there is merit to Hume’s position that we form “general ideas” of determinable classifications in close consideration of the particulars we experience and recall. Nevertheless, we may ask: is the significant use of a Humean general term parasitic upon its revival set, or is the revival set instead parasitic upon the determinable classification designated by a significant general term? Hume’s statements about whether or how a revival set underwrites the significant use of a general term are inconsistent. I shall argue that Hume’s contrasting statements about this crucial point reveal that the revival set is parasitic upon the determinable classification designated by that term, and that our terms for determinables can be used significantly—according to Hume himself—without appeal to any revival set, nor to any exemplary idea-object. This is because our imagination’s capacity to discriminate and to identify determinable resemblances does all the work required in and by Hume’s account of “general ideas” and of our significant, general use of terms.
On the one hand, Hume notes that a “word” recalls a particular idea-object, and its customary association with other resembling idea-objects (T 1.1.7.7, SBN 20). He comments on “the progress of the thought, either in reflection or conversation,” noting that:
Nothing is more admirable than the readiness with which the imagination suggests its ideas, and presents them at the very instant in which they become necessary or useful. (T 1.1.7.15, SBN 23)
However, in the second passage quoted at the beginning of §4, Hume recognises that we cannot literally recall all of the allegedly relevant instances we have experienced of idea-objects bearing the relevant resembling aspect. Speaking of individual members of the revival set of relevant idea-objects, Hume states:
They are not really and in fact present to the mind, but only in power; nor do we draw them all out distinctly in the imagination, but keep ourselves in a readiness to survey any of them, as we may be prompted by a present design or necessity. (T 1.1.7.7, SBN 20)
These two passages appear only three pages apart in Hume’s text. They would be inconsistent if, in the former, Hume meant by “ideas” what I have called ‘idea-objects’; they would be consistent if, in the former, Hume meant by “ideas” abstract or general ideas, so that the latter passage would indicate the irrelevance of any recollection of any member of a term’s associated idea-objects to cogent use of abstract ideas (or their associated general terms). The paragraph from which the first of these statements is taken, however, strongly suggests that its use of “ideas” is to be understood in the sense of ‘idea-objects’. This raises the question, which suffices for cogent general thought, the actual, or merely the potential, presence to mind of idea-objects within the revival set of idea-objects associated with any general term? I pose the question to dispose of it as irrelevant, because in either case, what supplies the relevant specificity to the thought or to the general term in use is the imagination’s classificatorily specific “resemblance,” whether determinate or determinable. This classificatory characteristic guides the imagination’s remarkable power to suggest relevant idea-objects, per the former passage. Per the latter passage, the revival set can be “present to the mind” merely “in power,” i.e., merely potentially, only because our imagination has developed the relevant, specific classification, which suffices unto itself for general thought and which alone enables us to recall (only) relevant instances, should we find need to do so, because the imagination’s classification specifies the relevance class, not vice versa. This case is typical, not exceptional. Though he insists that even quite abstract terms are learnt and are significant only because they are built upon the basis of our experience of many simple idea-objects, Hume recognises that we often don’t recall any idea-object whilst speaking, especially in speaking of such matters as “government, church, negotiation, [or] conquest” (T 1.1.7.14, SBN 23).Footnote 26 In this connection, Hume stresses that we do this with no risk of confusion (nor lapses) in our thinking. This entails that fleeting recollection of idea-objects associated with any significant term we use is irrelevant both to the significance and to our proper use of that general term, and to cogent thought using the “general idea” (not idea-object) in question, because what actually does the work of a Humean “general idea” is a specific classification of a relevant determinable resemblance among characteristics of particulars, developed by our imagination.Footnote 27
Hume insists that reasoning or understanding presupposes “clear ideas” or “conception[s]” (T 1.3.14.17, SBN 164).Footnote 28 Since such ideas or conceptions result from associations forged by the imagination among resembling idea-objects and impressions (T 1.1.3.4, 1.1.4.1, SBN 10–11)—however these resemblances are specified, as pertains to the topic under considerationFootnote 29—Hume would insist that he’s right to speak in the above passage of “imagination” rather than thought, reason or understanding. However, as noted above (§§3, 4), Hume’s actual account of our formation of significant words relies essentially on a range of intelligent acts of mind—expressly including reasoning, reflexion, acts of comparison, regarding or noticing—which are standardly ascribed to the understanding. These acts are underwritten in Hume’s account by specific “resemblances” distinguished and identified by our “imagination.” These “resemblances” or classifications may be perceptually specific or classificatorily specific, and in the latter case either determinate or determinable. Only classificatorily specific resemblances can account for distinctions of reason, on Hume’s account—indeed on any tenable account—of such distinctions. Since this discriminatory capacity of our imagination is one of its most general and established characteristics, this capacity belongs to understanding, on Hume’s account (see below, §6). Fortunately, the world happens to provide us with specific, recognisable regularities and differences among the qualities instanced by our impressions and idea-objects, which allow us to share a language, and to develop the variety of natural languages which largely share a common structure, Hume claims (T 1.1.4.1, SBN 10–11).Footnote 30 Hume’s inclusion of determinable resemblances within the imagination’s associative capacity has not been clearly discerned in the literature.Footnote 31 Further supporting evidence for Hume’s inclusion of determinable (and hence non-imagistic) classifications is provided in the following sections (§§6–8).
6. Hume’s Account of Distinctions of Reason
Hume recognises that his equation of distinctness and separability poses a serious question for his Separability Principle, namely, how to account for distinctions where no real separation is possible, e.g., between the figure and the colour of a marble sphere, or between motion and a body that moves (T 1.1.7.17–18, SBN 24–5). Only in this connection does Hume acknowledge that even simple idea-objects might contain “many different resemblances and relations” (T 1.1.7.18, SBN 25).Footnote 32 This is what I have called the perceptual specificity of idea-objects as determinate particulars (not determinables). Hume answers along expectable lines: When we see globes of different colours, and the same colour exhibited by differently shaped objects, we are able to notice their globular shape and to distinguish that shape from the various colours, even within any one coloured globe (T 1.1.7.18, SBN 25). “Able to notice” is not a prejudicial phrase, for Hume uses several such locutions:
[When] observing [sic] afterwards a globe of black marble and a cube of white, and comparing [sic] them with our former object, we find [sic] two separate resemblances, in what formerly seemed, and really is [sic], perfectly inseparable. After a little more practice [sic] of this kind, we begin to distinguish [sic] the figure from the colour by a distinction of reason; that is, we consider [sic] the figure and colour together, since they are, in effect, the same and undistinguishable; but still view [sic] them in different aspects, according to the resemblances of which they are susceptible. When we would consider [sic] only the figure of the globe of white marble we form [sic] in reality an idea both of the figure and colour, but tacitly carry our eye [sic] to its resemblance with the globe of black marble: and in the same manner, when we would consider [sic] its colour only, we turn our view [sic] to its resemblance with the cube of white marble. By this means we accompany our ideas with a kind of reflexion [sic], of which custom renders us, in a great measure, insensible. (T 1.1.7.18, SBN 25)
Throughout, Hume insists on the fully determinate character or perceptual specificity of individual objects, including idea-objects, whilst recognising that we can make fine-grained distinctions within any one impression or idea-object, among its various aspects, e.g., its shape and its colour. Once again, Hume’s account of how we can make distinctions of reason is replete with phrases for intelligent mental acts, concerning what we “distinguish,” “consider,” “view … in different aspects,” or “find” with some “practice”; how we “tacitly carry our eye” or “turn … our view” to some specific resemblance, in virtue of our “reflexion” on a quite specific feature of a plurality of idea-objects. Hume’s terms “eye” and “view” cannot be taken literally, because our literal eyes or our literal view only take in concrete, spatiotemporal individuals which exhibit perceptual specificity and so do not exhibit classificatory specificity. Plainly, Hume’s terms are metaphors for acts of attention, specifically, acts of discerning and attending to classificatorily specific features of concrete particulars, features which they can share with other individuals (cf. T 1.2.4.23–24, SBN 47). Hume’s account cannot dispense with these metaphors. Though they may again appear to invite the Standard Objection, the key point is rather that our capacity to attend to such fine, classificatorily specific discriminations is entirely due to our imaginative capacity to identify and to individuate classificatorily specific “resemblances” or characteristics shared among particular individuals (whether physical objects, impressions or idea-objects). Nothing less, and nothing more, than this is required for, or provided by, Hume’s account of this fine-grained associational capacity of human imagination. Note further that Hume’s discussion of his examples clearly indicates that the relevant classificatorily specific resemblances are not limited to determinate resemblances, but also include determinables, such as “colour” and “shape.”
According to Hume such classificatorily specific discriminations are cases of reasoning, because:
All kinds of reasoning consist in nothing but a comparison, and a discovery of those relations, either constant or inconstant, which two or more objects bear to each other. (T 1.3.2.2, SBN 73)
Hume’s definition of reasoning coincides with the definition of judgment in traditional term logic. This feature of Hume’s account of what he prefers to call “the imagination” cannot be discounted by noting that he later repudiates any such notion of judgment,Footnote 33 nor that much later he restricts “understanding” to “demonstrative and probable reasonings” (T 2.2.7.6 note, SBN 371 note). Exactly what enables us to associate resembling ideas, however fine-grained or determinable these discriminations may be, or to name these “resemblances,” Hume does not propose to explain. Presumably, these capacities are among the original powers of the soul, which Hume repeatedly (and understandably) insists are beyond human comprehension.Footnote 34 These “original powers” include the Copy Theory and Hume’s three basic laws of psychological association according to resemblance, contiguity and cause–effect.Footnote 35 We have now seen, however, that relevant “resemblances” run the gamut from the most concrete (e.g., duplicate particulars) to the most finely grained classifications, including determinables: e.g., “dusty periwinkle blue,” “shape,” “government” or “exist.”Footnote 36 These points are further corroborated by Hume’s accounts of the ideas of equality, space and time.
7. Hume’s Idea of Equality
Hume’s commitment to his objectualist account of thought, and the main defect of that account, are especially apparent in his account of the idea of equality. The idea of equality is a telling case for Hume’s Copy Theory and Concept Empiricism because equality is merely a relation between two or more things, and not a characteristic of any one thing unto itself. Hence the idea-object of equality copies no impression or set of impressions, but “arises merely from the comparison which the mind makes betwixt them” (T 1.2.4.21, SBN 46). Hume here admits that a “mere comparison” generates an idea-object! I say “idea-object” because this is Hume’s official view of “ideas,” of meaningful thoughts, and he speaks repeatedly of our “idea of equality,” though by the Copy Theory he should at most speak of our ‘equality-idea’. The main point, however, concerns Hume’s account of our significant use of the general term “equality,” in which he admits that, beyond whatever pair or group of particular perceptions are compared, no further idea-object plays any role. Once again, I attend to what Hume says about language and about our acts of attention and discrimination for what such remarks reveal about the functioning of our associative imagination, which enables us to use such language or to pay such discriminatory attention to particular perceptions.
Hume’s account of our comparison among objects regarding their quantitative equality or inequality draws the exact opposite conclusions from the same phenomena to which Socrates appealed when arguing that human thought transcends the senses. Socrates argued that our ability to recognize inequalities among sticks of similar lengths shows that we can identify both what equality is, and its absence among various physical objects of very nearly equal size.Footnote 37 Both Socrates and Hume recognise that the senses provide us only approximately equal objects, and cannot ground more than judgments of approximate equality. From these same premises, Socrates argues modus tollens tollendo in favour of our power of rational conception, whilst Hume argues modus ponens ponendo against our having any such powers. Where Socrates concludes that the powers of the mind exceed those of the senses, because the mind can grasp equality itself, whilst the senses can only present us approximately equal things; Hume concludes that our power of conception is hostage to the senses, and cannot have any equality-idea-object more exact than our senses and sensible standards or measuring instruments afford. This is the nub of Hume’s basis for criticising mathematical knowledge and reasoning as lacking the precision and certainty it is pretended to have (T 1.2.4.30, SBN 50). Though Hume tried to develop and to defend his criticism of mathematical knowledge to the end of his life, the better wisdom has always been that of Lord Stanthorpe, that Hume’s critique of geometry was defective and should not be published.Footnote 38
Hume’s criticism of mathematics rests squarely on his objectualist account of thought, and is relevant here as evidence of how deeply committed Hume was to that account, and as evidence that even in developing that account, Hume was sufficiently acute to recognise that his objectualist account of thought was inadequate, because the very idea-object of equality is not, and cannot be, produced by impressions, whether singly or collectively.Footnote 39 Instead, as Hume insists, this idea-object results only (“merely”) from the comparison the mind makes among the sizes of particular objects. It has become common recently to say, e.g., that our equality-idea is caused to occur within our Humean bundle-minds by imaginative association among (nearly) equal impressions or idea-objects. If so, this kind of causality differs from that allowed by the Copy Theory; this difference has not been adequately acknowledged by the proponents of such causation. Neither has it been adequately acknowledged that the suggested causal relations are so vaguely described that they are merely posited if and when a Humean idea-object is said to occur, though it does not conform to the Copy Theory (nor to Concept Empiricism), and it violates Hume’s basic principle that any and every perception is determinate, which is central to his criticism and rejection of Locke’s account of abstract ideas. This is both poor empiricism and naturalistic representation on the philosophical cheap; it may serve to deflect unwelcome objections to Hume’s views, but it provides no corresponding insight into Hume’s views nor into the phenomenon of the generality of thought. Representation must be earned, and Hume took this desideratum seriously.Footnote 40 Hume’s Copy Theory, Concept Empiricism, three psychological laws of association (including his official empiricist analysis of causality as 1:1 correlation of instances of event types with spatiotemporal contiguity) might, at best, be necessary conditions for defining concepts or terms, but they are not sufficient conditions for specifying, nor for generating, the conceptual content of determinable concepts.
The equality-idea-object cannot result from impressions in accord with the Copy Theory. Accordingly, the “comparison” of which Hume speaks must be understood in terms of his account of abstract ideas and general terms. Hence there is and can be no basis within Hume’s philosophy for calling the idea of equality an “idea-object.” To be aware of the equality of any two objects is to be aware that those two objects are equal. Likewise, to be aware of the inequality of any two objects is to be aware that those two objects are unequal. This explicit awareness of equality (or of inequality) requires grasp and use of a classification—not an occurrent idea-object—of equality, and of its instantiation (or lack thereof) by any plurality of items in view. Once again, I do not base my case on Hume’s account of language, nor on his locutions for mental acts, but rather on the capacious discriminatory power of our imagination which alone (on Hume’s account) enables us to associate, e.g., any pair of particulars as equal or unequal in any specific dimension. This capacity of our imagination alone underwrites the kinds of awareness, consideration or comparison of which Hume writes. If salient resemblances include equality and inequality among particulars, this requires admitting that “inequality” is a determinable, not a determinate, species of resemblance. Here, too, Hume’s official empiricist tenets—the Copy Theory, Concept Empiricism and the three official laws of association—fail to explain the significant use of general terms solely on the basis of the occurrence of particular perceptions, including “idea-objects,” within any one of us qua bundle of (concrete, determinate) perceptions. To suggest that an idea-object (whether secondary, tertiary or n-ary) copies an act of comparison among two or more particular perceptions regarding their equality or inequality, requires distinguishing between intelligent acts and their objects strictly parallel to the Cartesian distinction between the formal and the objective reality of “ideas” which Hume expressly rejected (above, §2). Hume’s official objectualist account of thought falters here. Instead, a crucial feature of our prevalent use of determinable classifications is provided solely by our imagination’s capacity to discriminate and to identify determinable classifications.
Using concepts to represent various objects or events or their features or relations is yet a further, so far unexplained, capacity: officially, Humean “names” can only be yet further auditory, visual or (in the case, e.g., of Braille) tactile idea-objects, which do not at all “resemble” the idea-objects they purportedly designate. Here, too, Hume can only appeal to the wondrous powers of human imagination.
8. Hume’s Ideas of Space and Time
This same conclusion results from reconsidering Hume’s accounts of our ideas of space and time. Hume opens his account of the idea-objects of space and time by reiterating his principle that idea-objects are nothing but copies of prior impressions (T 1.2.3.1, cf. 1.2.3.10, SBN 33, cf. 37). He recognises that only impressions of sense pertain to the idea-object of space, because impressions of reflection only concern “our passions, emotions, desires, and aversions,” which are not spatial (T 1.2.3.17, SBN 33). The problem with appealing to impressions or idea-objects of sense in this context is very like the problem Hume confronted in the idea of equality: the idea of space space and, even more obviously the idea of time, cannot be reduced to any one, nor to any group, of idea-objects. Time and space or extension (Hume equates the latter pairFootnote 41) are manifest in specific relations among (or sometimes within) idea-objects, in the “manner” in which the relevant impressions or idea-objects appear.Footnote 42 Hume expressly insists that the idea-objects of space or of time are neither impressions nor idea-objects in addition to the spatially or temporally arrayed idea-objects we consider (or rather, which occur within each of our bundles of perceptions; T 1.2.3.4, 1.2.3.10, SBN 34, 36). Instead, the idea-object of space is generated in accord with Hume’s account of abstract ideas. When we find, Hume writes,
… a resemblance in the [spatial] disposition of coloured points, of which [extended things] are composed, we omit the peculiarities of colour, as far as possible, and found an abstract idea merely on that disposition of points, or manner of appearance, in which they agree. (T 1.2.3.5, SBN 34)
Expressly, our “abstract idea” of space is “found[ed] … merely on that [spatial] disposition of points,” or the spatial “manner of [their] appearance.” Indeed so: our idea of space is significant even though it copies no impression nor idea-object, nor any plurality thereof. It is significant because it selectively identifies the spatial configuration of objects (impressions or idea-objects) or their features.
This way of putting Hume’s point obscures some important details. His Copy Theory and Concept Empiricism each entail that, for any case of any one extended impression or idea-object, or likewise for any case of any one spatial array of impressions or idea-objects, we can at most obtain only one idea-object (or particular determinate concept), the content of which is nothing other than the determinate spatial configuration of that one extended particular perception (or that one spatial array), which is (or is tantamount to) its determinate spatial array of coloured (or tactile) points.Footnote 43 Taken by itself, any such idea-object (or particular determinate concept) represents only that one array; call it ‘Extensional Array No. 1’. This process, this generation of any determinate extended idea-object or of any determinate classification of any distinct, determinate, individual Extensional Array, may be repeated ad libitum, to produce however plentiful a series one can distinguish and (re)identify. However, such determinate Extensional Array idea-objects or determinate Extensional Array classifications (names), are, each of them, determinate and also determinately distinct. Each can only serve to identify (or to classify or to name) exact replicas of the same determinate extensional array. Neither singly, nor taken together, can they provide an abstract idea of space, nor of spatial nor of extended because these are determinables.Footnote 44 Neither Hume’s Copy Theory nor Concept Empiricism can provide, define or specify determinables.Footnote 45 Only the determinable “spatial” or “extended” provides the required generality; it alone can associate (or classify) Extensional Arrays Nos. 1 to (say) 2,398 as resembling each other spatially. This generality, which is so fundamental to human thought, is provided within Hume’s account solely by the capacious discriminatory powers of our imagination, which can identify not only determinate classifications (specific, determinate resemblances, including determinate extensions: specific sizes, lengths or arrays), but also determinable classifications (“abstract” resemblances, so to speak), such as spatiality.
The case is even stronger and more clear-cut regarding the idea-object of time. The essence of time, Hume writes, is that its “parts” are successive, never concurrent (T 1.2.2.4, 1.2.3.8, SBN 31, 36). Hume contends that our idea of time is derived from all kinds of successive perceptions (T 1.2.3.6–7, SBN 34–5). For example, upon hearing five successive notes from a flute, our mind “only takes notice [sic] of the manner in which the different sounds make their appearance,” and on that basis generates an abstract time-idea, which can then be used on other occasions (T 1.2.3.10, SBN 37). Once again, according to the Copy Theory and to Concept Empiricism, any such temporal manner-idea-object can only be used (rather than misused) to classify exact replica determinate periods or durations or intervals. Generating arbitrarily many such determinate temporal manner-idea-objects cannot account for the generality—that is, for the determinability—Hume recognises is required for an idea of “time” (or better, as I shall say, of temporality). Once again, only by attending selectively to—by taking “notice” of—the determinable temporal “manner” of the appearance of several distinct, determinate temporal-sequence-idea-objects can we generate a significant use of the general term ‘temporal’. This determinable classification is generated, on Hume’s account, only by our capacious, discriminating, also generalising imagination.
Only this capacity of our imagination allows Hume to deny, as he expressly and rightly does, that our ideas of space and of time are distinct idea-objects:
The ideas of space and time are, therefore, no separate or distinct ideas [sic], but merely those of [sic] the manner or order in which objects exist …. (T 1.2.4.1, SBN 39)
Officially ideas, as particular perceptions (i.e., idea-objects), are mutually distinct, separate (or at least separable, if they are contiguous) and fully determinate. Hume recognises here that idea-objects cannot account for our significant use of the terms nor of the “general ideas” (i.e., determinable classifications) of spatiality or of temporality (cf. T 1.2.5.28–29, SBN 64–5). Notice that Hume’s phrase “idea … of the manner or order in which objects exist” is correct, where “of” must be understood intensionally; substituting “idea-object” or more specifically “manner-idea” into Hume’s statement reveals the inadequacy of Hume’s official, objectual account of thought, of its supporting Copy Theory and also of Concept Empiricism.
The case of time and temporal succession is more complicated than that of space, though also more revealing philosophically. Because the moments of time are successive, to be aware of the temporal manner in which a series of idea-objects appears, such as Hume’s five successive notes of a flute, requires that we recall the prior notes and their temporal intervals as we hear each successor. Developing or defining the idea of temporality (time) requires memory. In stark contrast to the imagination, our memory has no power to alter, but only to preserve the order and position in which impressions originally occurred (T 1.1.3.2–3, 1.3.5.3, SBN 9, 85).Footnote 46 It is important to stress, however, what Hume occasionally notes expressly, that the relevant “preservation” of the order and position of past impressions involves making us aware of their order: memory represents past impressions as past; it informs us about the past.Footnote 47 If this is a brute fact about human memory, then Hume provides no account of this capacity of memory, but merely describes and appeals to it. The preceding discussion, however, shows that this capacity of human memory need not be taken by Hume as a brute fact. Instead, Hume can appeal to his account of the discriminatory capacity of human imagination, in conjunction with memory, to distinguish and to identify the classificatorily specific though determinable “resemblance” of “being determinately sequenced temporally.” Were our imagination (together with memory) incapable of such a classificatorily specific, yet determinable, discrimination of temporal characteristics generally, we could not (on Hume’s account) identify or be expressly aware of past impressions or idea-objects as past, and so as temporal.Footnote 48
For the same reasons considered with respect to spatiality, the Copy Theory and Concept Empiricism can account only for specific, determinate idea-objects or specific names for each member of any plurality of determinate temporal sequence-idea-objects. Also with respect to the determinable classification “temporal” (or “temporality”), only the capacity of our imagination (cum understanding) to discriminate classificatorily specific and yet determinable characteristics, aspects or “resemblances” among particulars enables us to have and to use a “general idea” (not an idea-object) to think—or a general term to discourse—about temporal aspects of various determinate sequences of idea-objects or impressions. Within Hume’s account, only our imagination can or does provide for properly general thought by providing for classificatorily specific yet determinable “resemblances,” of any sort—of any sortal—which does not require strict replication of any one determinate instance of a distinguished “resemblance.” Once any such determinable “resemblance” (or classification) is distinguished and identified (solely) by the imagination, its further use requires no revival set of idea-objects. Any relevant revival set of idea-objects can only be specified by that determinable resemblance cum classification, not vice versa.
9. An Objection?
Hume is adamant that the mind cannot,
by revolving over a thousand times all its ideas of sensation, ever extract from them any new original idea, unless nature has so framed its faculties, that it feels some new original impression arise from such a contemplation. (T 1.2.3.10, SBN 37)
This prompts the question: How exactly has nature framed our capacities, and might they be so framed on Hume’s account as to avoid or to rebut the present analysis? I have stressed that to account for the generality of thought by accounting for, or at least giving due acknowledgement to, the role of determinable classifications (however specific or “abstract”), Hume’s sole recourse is to the capacity of human imagination to distinguish and to identify determinable classifications. If this is so, then the above remark can be understood as noting, if unwittingly, the limits of both the Copy Theory and of Concept Empiricism to account for determinable classifications.
How then might Hume’s Copy Theory or Concept Empiricism be kept alive? One suggestion would be to claim that Hume’s Copy Theory allows our forming an idea-copy of, e.g., just the temporal manner of the succession of Hume’s five notes, or of just the spatial manner in which coloured or tactile points are arrayed in some spatially extended impression or idea-object (or plurality thereof), so that these special idea-objects can be, or can serve as, respectively, the abstract idea of time or of space.Footnote 49 However the details of this suggestion might be worked out, it confronts two crippling difficulties. Although the indefinite English term “copy” may be more accommodating to the suggestion here considered, insofar as Hume’s Copy Theory is taken strictly,Footnote 50 the suggested abstract idea-copies would be, respectively, an idea-object of just this determinate temporal sequence (which happens to be instantiated by those five notes), or of just this determinate spatial array (which happens to be instantiated by those coloured or tactile points). Each resulting abstract idea-object remains specific and determinate, and can only classify exact replicas. These abstract though determinate idea-objects can neither be, nor can they generate, an abstract though determinable temporality-idea-object, nor an abstract though determinable spatiality-idea-object. For the suggested abstract though determinate idea-objects to be used generally requires the imagination to generate the required determinable classifications or “general ideas,” ‘temporal’ and ‘spatial’ (to which we can then, as Hume says, assign a name; we need not now ask how). The suggested abstract idea-objects provide no advance upon Hume’s own analysis.
It is not tendentious in this connection to distinguish between the suggested idea-objects and what I have just referred to as “Hume’s own analysis.” The suggested “copies” of any one determinate spatial or temporal array, as such, are especially abstract because they omit any and all further characteristics of the impressions or idea-objects originally so arrayed. Consider again how Hume describes our development of an abstract idea of space:
[Upon] finding a resemblance in the [spatial] disposition of coloured points, of which [extended things] are composed, we omit the peculiarities of colour, as far as possible, and found an abstract idea merely on that disposition of points, or manner of appearance, in which they agree. (T 1.2.3.5, SBN 34)
Hume does not write, “we omit the peculiarities of colour,” but carefully qualifies this by adding that we omit them “as far as possible.” Uncoloured, non-tactile points simply are not particular perceptions, on Hume’s view, nor on any plausibly human view of sensory awareness; neither is an uninstantiated array. Consequently, they cannot occur within a Humean mind qua bundle of impressions and idea-objects (perceptions all). The suggested abstract copy-ideas are “especially” abstract because they are entirely abstracted from any sensory basis. This Hume’s empiricism cannot and does not countenance; the suggested “copies” violate Hume’s Copy Theory.Footnote 51 Finally, the suggested especially abstract ideas require neglecting Hume’s express claims, quoted above, that no further abstract idea-object is involved in forming the abstract ideas of space or of time, beyond the original sensed array of sounds or coloured (or tactile) points. Because the suggested especially abstract copy-idea-objects sever these basic links with Hume’s sensationist philosophy of mind, they are to no avail.
This problem is remedied, if at all, not by making Hume’s Copy Theory ever more abstract, but by appealing instead to Concept Empiricism, which provides for naming specific, determinate arrays of impressions or idea-objects, whether perceived or recombined at will. Does naming just the manner in which five notes appear, help to show that Hume can account for determinable classifications according only to his Copy Theory and Concept Empiricism? This question can be answered without wading into difficult issues about whether or how Hume can account for “words” and their significance within his bundle theory of mind, and the official principles of psychological association which govern their bundling. Per above (§8), Concept Empiricism can only account for names of specific, determinate instances of some quantitatively or qualitatively determinate “resemblance.” Concept Empiricism is limited to naming exact, determinate replicas or their recombinations, and cannot account for the proper generality of thought, nor of Hume’s “general ideas,” which are constituted by determinable classifications (§5).
Trying to make Hume’s Copy Theory more robust (or: less committal) than it initially appears to be cannot foreclose the analysis developed here, according to which Hume’s account of the determinable classifications required for the generality of human thought and for “general ideas” is entirely due to our imagination’s capacity to discriminate and to identify classificatorily specific, fine-grained though determinable classifications. According to Hume, determinable classifications, too, are among the “resemblances” discriminated by our imagination. Because making discriminations is one of the most “general and … establish’d properties of the imagination” (T 1.4.7.7, SBN 267), if not the foremost such property, this crucial capacity is due to human understanding. This holds of classifications of, or “resemblances” among, determinate perceptually specific replicas, determinate classificatorily specific replicas (however selective or fine-grained) and also of determinable classifications, of just the kind Hume stresses in connection with “general ideas” (above, §5).
10. Some Critical Conclusions
I conclude that Hume’s accounts of the abstract or “general” ideas of existence, space, time and equality, like his account of “distinctions of reason,” all of which involve determinables, are anomalies to his Copy Theory and to Concept Empiricism. However, they are not anomalies to Hume’s account of the generality of thought, because the required generality, involving determinable classifications, is due to our capacious imagination cum understanding. Hume’s accounts of these abstract, general ideas ultimately all point to the same conclusion: idea-objects are both insufficient for, and irrelevant to, Hume’s account of abstract determinable “general” ideas and to our reliable and significant general use of terms. All the philosophical burden of Hume’s account of the generality of thought is instead borne by our imaginative, that is, our intellectual capacity (of understanding) to discriminate and to identify classificatorily specific, fine-grained yet determinable “resemblances” among characteristics shared by particular impressions or idea-objects.Footnote 52
Hume summarises his system in one sentence:
According to my system, all reasonings are nothing but the effects of custom, and custom has no influence, but by enlivening the imagination, and giving us a strong conception of any object. (T 1.3.13.11, SBN 149)
The main problem with Hume’s system, so summarised, is that it sweeps under the rhetorical rug the fact that core instances of the relevant “customs” or likewise habits (of thought) are developed, defined and constituted by our determinable classifications, all of which are ultimately due to our understanding and its discriminatory judgments. Hume can insist that judgment is based on custom (T 1.3.13.8, SBN 147) only if he either acknowledges or disregards the basic comparative judgments among impressions or idea-objects, stressed in his account of abstract ideas and ascribed to “reasoning” and to “understanding” (T 1.3.2.2, SBN 73), which alone enable us to identify any salient determinable classifications of impressions and idea-objects, which alone enable us to develop the significant use of any general term (the generality of which is greater than merely naming determinate replicas, however finely or densely specified). Without general terms, we would be incapable of either reflection or discourse, including those found throughout Hume’s Treatise.Footnote 53
Hume’s official empiricist views, the Copy Theory and Concept Empiricism, are insufficient to account for our general ideas, for our significant use of general terms and for our imagination’s—or rather our understanding’s—discriminatory capacities to identify determinable classifications. Consequently, Hume’s official objectual account of thought is, by his own discerning analysis of the generality of thought, inadequate. By the same token, Hume lacks any specifically empiricist explanation or account of our generation and use of determinable classifications, though he is right to stress that we develop such classifications through close consideration of whatever features the particulars we experience present to us. These important points have been neglected by Hume’s expositors and critics alike.Footnote 54 What makes Hume’s Treatise so much more deeply philosophical than his Enquiries is that in the Treatise Hume’s enormous acuity and honesty enable him to recognise the limits of his official empiricist principles and to continue philosophising about the remaining, as yet unaccounted doxastic phenomena, in ways which often reveal precisely which additional principles, which supplements to his official empiricism, are required to account for these further phenomena. Hence it is unfortunate that Hume’s first Enquiry, rather than the Treatise, became the basis of so much of twentieth-century empiricism. This broad theme cannot be explored here. Suffice it to recall that, as recently as 1984, Donald Davidson found it necessary to remind philosophers of the important distinction between a shirt being green, and the thought that it is green.Footnote 55
