Locke, John (1632–1704)

5. Substances, mixed modes and the improvement of language

On Locke’s account of communication (Essay III.i–ii), names should, by common convention or special agreement, excite in the hearer’s mind just the same ideas as they are associated with in the speaker’s mind. Collaborative progress in the sciences depends on ‘clear and distinct’ or ‘determined and determinate’ ideas – that is, on consistent and agreed association of ideas and words (Essay II.xxix; compare ‘Epistle to the Reader’). Locke’s discussion of language is shaped by his belief that these conditions of the transference of knowledge were in his time commonly unsatisfied, especially in two domains. First, there was no agreed classification of ‘substances’ (living things and chemicals) based on careful observation and experiment. Second, the ideas associated with the names of mixed modes often varied both in the usage of different people and in that of the same person at different times. Two mistakes in particular disguise these shortcomings of language. The first is the assumption that a common set of words ensures a common language in the full sense, with a shared set of meanings. So people may argue about ‘honour’ and ‘courage’ without realizing that they mean different things, or nothing at all, by the words. The second mistaken assumption is that words have meaning by standing for things directly, as if the meaning of ‘salt’, ‘gold’ or ‘fish’ were fixed demonstratively, by what is named. The first assumption chiefly corrupts our thought about mixed modes, the second relates ‘more particularly, to substances, and their names’ (Essay III.ii.5). Locke’s radical and influential views about the latter will be considered first.

The ‘idea of substance in general’ employed in ideas of specific substances is the idea of something unknown underlying the attributes known by experience (see §4 of this entry). Many have objected, following Leibniz, that here Locke confusedly postulates ignorance of the subject of attributes which is not ignorance of attributes of the subject. Yet he holds that our ignorance of ‘the substance of body’ and ‘the substance of spirit’ is an ignorance of the natures of these things – ignorance manifested in our inability to understand the internal cohesion or (he adds in later writings) mutual attraction of bodies, or to explain what thinks in us, and how it does so. His approval of the corpuscularian hypothesis and Newton’s mechanics is qualified – the best available physical theory leaves too much unexplained to be the whole truth (Newton did not disagree). The idea of substance is a place-marker for essences which are unknown, but knowable, if possibly not by human beings (see Newton, I.).

One feature of Locke’s theory which has made difficulties for the present interpretation is the distinction he makes between substance and ‘real essence’. The real essence of a thing, Locke says, may be taken for ‘the very being of a thing, whereby it is what it is’, the ‘real internal, but generally in substances, unknown constitution of things, whereon their discoverable qualities depend’ (Essay III.iii.15). Nevertheless, ‘essence, in the ordinary use of the word, relates to sorts’ (Essay III.vi.4). Species and genera, or sorts of things, Locke asserts, are creatures of the understanding, with membership determined by abstract ideas made on the basis of experienced resemblances, not by the presence in each of a specific form, or by a common derivation from a divine archetype. Ultimately it is a matter of arbitrary definition which observable resemblances we count as necessary for membership of this or that named sort. It is not just that specific real essences are unknown, since (Locke argues) even if we did know the real constitution of things as well as clock-makers know the works of clocks, it would still be up to us where to draw the boundaries between species, and what to include in our abstract ideas or ‘nominal essences’. The real essence of a species can therefore only be ‘that real constitution…which is the foundation of all those properties, that are combined in, and are constantly found to co-exist with, the nominal essence’ of the species (Essay III.vi.6). Here, the model is that of a universal matter determinately modified as a variety of particles interacting mechanically so as to constitute the material things of ordinary experience. Since at the fundamental level these observable quasi-machines differ from one another merely quantitatively, and can do so by indefinite degrees, there are no absolute boundaries among them. There are only the discernible resemblances and differences consequent on their underlying mechanical differences – ‘the wheels, or springs…within’. Even more certainly our actual classification is not based on knowledge of any such boundaries. Talk of the real essence of a species, and the distinction between its ‘properties’ and its ‘accidents’ (’properties’ flowing from the essence), are therefore, contrary to Aristotelian assumptions, de dicto and relative to the nominal essence defining the name of that species (see Aristotle §8; De re/de dicto; Essentialism).

This conception of a real essence assigns it a role closely related to that of substance. What, after all, is the ‘unknown cause of the union’ of any of the ‘combinations of simple ideas’ by which ‘we represent particular sorts of substances to ourselves’, if not the real essence underlying the nominal essence in question? Yet Locke sometimes distinguishes both the notion and knowledge of real essence from the notion and knowledge of substance. That is not, however, because the ‘substance’ is an irremediably unknown subject underlying even essence, but because it is the common stuff of a variety of species of things, ‘as a tree and a pebble, being in the same sense body, and agreeing in the common nature of body, differ only in a bare modification of that common matter’ (Essay II.xiii.18). The unknown modification is the specific ‘real essence’, and the equally unknown general nature of matter is the ‘substance’. Locke also envisages deeper differences of kind between substances: ‘God, spirits and body’ are all ‘substances’ only because we think of each of them indeterminately as something, not because of a shared nature. But by the same token we distinguish spirit and body only because we cannot understand how matter could think, not because we can grasp their separate essences, as Descartes had supposed (see Descartes, R. §8). Indeed, since we are equally unable to understand how spirit and body might interact, or how spirits could occupy places, the issue between materialist and immaterialist accounts of minds is for Locke undecidable, and at best a matter for speculation.

Locke’s corpuscularian conception of a world of machines, resembling and differing from one another by continuous degrees, is consonant with his independent epistemological conviction that names have meaning only through association with ‘ideas’, rather than directly with ‘things as really they are’. Together they motivate his programme for improving natural classification which advocates, not the allegedly impossible Aristotelian ideal of identifying the natural hierarchy of genera and species, but general agreement on a practically useful way of gathering and ordering the things in the world, taking into account such dependable concomitances of qualities and powers as appear to careful observation and experiment. Locke saw the future of biology and chemistry – and even of mechanics – in descriptive ‘natural history’, justifiable as a useful, orderly record of dependable means to ends but falling short of systematic ‘science’. Despite its apparent pessimism, his view has survived in biological taxonomy as a continuous tradition of scepticism as to the reality of our taxonomical divisions. In semantic theory, Locke’s broad conception of how the names of substances have meaning has only recently been eclipsed by a quasi-Aristotelian view (see Kripke, S.A.; Putnam, H.).

Locke saw equal need for a programme of agreeing definitions in ethics, where his target is less the notion that moral and political terms name independent realities, than the assumption that the very existence of a word in a language ensures that it has a fixed, common meaning. ‘Common use’, Locke concedes, ‘regulates the meaning of words pretty well for common conversation’ (Essay III.ix.8) – for the ‘civil’ rather than ‘philosophical’ use of words. But where precision is required, as in the establishment or interpretation of a law or moral rule, reliance on ordinary usage leaves us vulnerable to the trickery of rhetoricians who prove bad qualities good by shifting the meaning of terms; or to the subtleties of interpretors, whether of civil or revealed law, who render unintelligible what started off plain. The remedy is to give the names of virtues and vices, and of social actions, roles and relations the fixed and unequivocal definitions necessary for a clear and unwavering view of right and wrong, ‘the conformity or disagreement of our actions to some law’.