Locke, John (1632–1704)

6. Knowledge and belief

Like Platonists and Cartesians, Locke drew a strong distinction between knowledge and belief (also called opinion, judgment, or assent), but the ground and placing of this division between two forms of propositional ‘affirmation’ differed from theirs fundamentally. As the distinction is expounded in Book IV of the Essay, universal knowledge or ‘science’ does not have special objects, whether a transcendent intelligible world in the mind of God or innate intellectual ideas. Its difference from general belief lies in the way in which ideas are related in the mind. In universal knowledge, the ‘connection and agreement, or disagreement and repugnancy’ of ideas is ‘perceived’, whereas in belief it is ‘presumed’ on the basis of ‘something extraneous to the thing I believe’. To follow a proof gives knowledge of the conclusion, whereas to accept the conclusion on the authority of a mathematician constitutes only belief. Similarly in the case of ‘sensitive knowledge’ of particular existence, what we ourselves perceive we know to be so, but what we infer, or accept on testimony, we merely believe.

Knowledge, as well as assent, is subject to ‘degrees’: there are degrees, not only of probability, but of ‘evidence’. The first degree of knowledge is intuitive knowledge, in which the mind ‘perceives the truth, as the eye doth light, only by being directed toward it’. Intuition ‘leaves no room for hesitation, doubt, or examination’. The second degree is demonstrative knowledge, where the truth is perceived by the aid of one or a chain of ‘intermediate ideas’. Doubt or mistake is possible at any point in the sequence with respect to connections not currently in view. Hence, ‘Men embrace often falsehoods for demonstration’. Locke’s chief model for ‘intermediate ideas’ is geometrical: for example, the lines employed in the intermediate steps of the Euclidean proof which allow us to see that the angles of a triangle are equal to the angle on a straight line. Although his conception of intuition can seem Cartesian, the profound difference is that, for Locke, ideas which are objects of intuition are essentially a product of sense (including reflection) and imagination. As Draft B puts it, the angles and figures I contemplate may be ‘drawn upon paper, carved in marble, or only fancied in my understanding’ (Drafts vol. 1: 152). Consequently Locke often talks as if we can literally perceive a necessary relation between ideas. Another difference from Descartes, as also from Hobbes, is that he rejects the pretensions of proposed analytical methods to uncover self-evident principles from which the phenomena can be deduced.

The third degree of knowledge is sensitive knowledge of the existence or ‘co-existence’ of qualities in external things. Locke’s first introduction of this category seems tentative, even an afterthought, as if it is called knowledge only by courtesy. In order to fit his main definition of knowledge it has to be interpreted as the perception of the agreement of ideas of sensible qualities with the idea of existence, an analysis Locke unsurprisingly declines to develop. Yet ‘sensitive knowledge of existence’ does straightforwardly satisfy his other definition of knowledge: what is known in sensitive knowledge (that is, that something external is causing an idea of sense) is known directly, ‘perceived’ and not inferred (see §2 of this entry). Locke was writing in a context in which, despite Gassendi’s Epicurean claim that sensory knowledge is the most evident of all, it was widely assumed that knowledge in the full sense comprises only knowledge of necessary first principles, demonstrated ‘science’, and perhaps reflexive knowledge. Locke wanted both to concede to orthodoxy that the evidence and certainty of our sensory knowledge is not as high as that of intuition and demonstration, and to insist that, nevertheless, ‘sensitive knowledge of existence’ does give a degree of immediate certainty and ‘deserves the name of knowledge’.

Knowledge is also categorized in terms of four propositional relations (forms of ‘agreement’) between ideas, namely ‘identity (or diversity)’, ‘relation’, ‘necessary connection or coexistence’ and ‘existence’ (Essay IV.i). By ‘identity’ Locke intends tautologies such as ‘Gold is gold’ and ‘Red is not blue’. Intuitive knowledge of such identities is achieved simply by discerning ideas. The category also includes such truths as ‘Gold is a metal’ or ‘Gold is malleable’, when the property predicated is included in the thinker’s definition of gold. Thus ‘identity’ covers all and only ‘trifling’ or ‘verbal’ propositions (see §2 of this entry).

The categories, ‘relation’, ‘necessary connection or coexistence’, and ‘existence’, on the other hand, together include all ‘instructive’ propositions. The category, ‘relation’, in part a response to Locke’s earlier difficulty over the informativeness of mathematics (see §2 of this entry), also marks his rejection of analytical methods in science. As well as geometrical axioms and theorems, ‘relation’ presumably includes more exciting Lockean principles: as that, if anything changes, something must have a power to make it change; that, if anything exists, something must have existed from eternity; and that a maker has rights over his artefact. Categorical propositions about natural things, however, fall either under ‘existence’ or under ‘necessary connection or coexistence’. Our own existence is known intuitively, God’s existence demonstratively (Locke employs an idiosyncratic hybrid of the cosmological and teleological proofs), and, as discussed, the existence of bodies by sense. The category ‘necessary connection or coexistence’ owes its disjunctive name to a rather complicated relation between particular and universal propositions. Particular coexistences are perceived by sense, for example, when we simply observe that yellowness, heaviness and the metallic qualities coexist in a particular subject together with malleability (that is, that this gold is malleable), without perceiving necessary connections between them. Locke assumes, however, with most mechanists, that necessary connections do hold between universally coexistent properties even if we cannot perceive or grasp them. Since he contends that no natural science based on the essences of substances has been achieved, he offers only very limited examples of perceived necessary connections, as ‘whatever is solid is impenetrable’ and ‘a body struck by another will move’. (According to the short, posthumously published Elements of Natural Philosophy, the laws of inertia are evidently necessary, but the law of gravity is based only on experience.) In the absence of knowledge, beliefs in universal coexistences (for example, that all gold dissolves in aqua regia), when we presume unperceived connections, may be inductively based on sensitive knowledge of particular coexistences. That is descriptive natural ‘history’, not ‘science’. In general, if the idea of a particular quality is deducible from the idea of a substance, that is only because the predication of that quality is an identity: that is, universal propositions about substances, if certain, are ‘trifling’ and, if ‘instructive’, are uncertain (see §2 of this entry). In contrast, instructive a priori sciences are possible just because their objects are constructed by us: our ideas of simple or mixed modes, formed without essential reference to actuality, themselves constitute the subject-matter of mathematics and ethics. In other words, these demonstrative sciences are possible, as natural science is not, just because they deal, hypothetically, with abstractions.

The degrees of assent are ‘Belief, Conjecture, Guess, Doubt, Wavering, Distrust, Disbelief, etc.’ (Essay IV.xvi). Probability is ‘the measure whereby [the] several degrees [of assent] are, or ought to be regulated’. When assent is unreasonable, it constitutes ‘error’. Reasonable assent is regulated according to the proposition’s conformity with the thinker’s own experience or the testimony of others. The proposition may concern ‘matters of fact’ falling within human experience, or else unobservables lying ‘beyond the discovery of our senses’. Locke identifies four broad degrees of probability with respect to ‘matters of fact’: (1) when the general consent of others concurs with the subject’s constant experience; (2) when experience and testimony suggest that something is so for the most part; (3) when unsuspected witnesses report what experience allows might as well be so as not; and (4) when ‘the reports of history and witnesses clash with the ordinary course of nature, or with one another’ – a situation in which there are no ‘precise rules’ for assessing probability. Finally, with respect to unobservables, ‘a wary reasoning from analogy’ with what falls within our experience ‘is the only [natural] help we have’ and the only ground of probability (see Descartes, R. §4). Although Locke, in striking contrast to Descartes, brings probability into the centre of epistemology, ‘belief’ is always treated as a practical surrogate for ‘knowledge’, and he takes induction itself to be grounded on the assumption of underlying, unknown necessary connections: ‘For what our own and other men’s constant observation has found always to be after the same manner, that we with reason conclude to be the effects of steady and regular causes, though they come not within the reach of our knowledge’ (Essay IV.xvi.6).

Another deliberate and radical difference from Descartes relates to the role of will in cognition. For Locke, knowledge is like sense perception: we may choose where and how hard to look, but we cannot then choose what we see. Belief is similar: ‘Assent is no more in our power than knowledge…. And what upon full examination I find the most probable, I cannot deny my assent to’ (Essay IV.xx.16). Yet we are morally responsible for both error and ignorance in so far as it results from our not employing our faculties as we should. In a number of chapters of the Essay, Locke examines the causes of error, finding them, with many writers of his time, in the same appetites, interests, passions, wayward imaginings and associations of ideas as may motivate voluntary actions. Linguistic confusion, and its deliberate exploitation (see §5 of this entry), sometimes plays a role, and sometimes, like Malebranche and others, Locke has direct recourse to physiology, explicitly merging his explanations of error with explanations of madness. In contrast to Hobbes, he places the merely habitual ‘association of ideas’ in the pathology of ‘extravagent’ thought and action: ‘all which seems to be but trains of motion in the animal spirits, which…continue on in the same steps…which by often treading are worn into a smooth path, and the motion in it becomes easy and as it were natural’ (Essay II.xxxiii.16). But culpable error arises, on Locke’s official view, when we ‘hinder both knowledge and assent, by stopping our enquiry, and not employing our faculties in the search of any truth’ (Essay IV.xx.16). It is the failure to use our power to pause for ‘full examination’ which leaves a space for beliefs motivated by interests and passions. But this two-stage model – the first stage voluntary, the second involuntary – proves too difficult to maintain, and sometimes passions and interests are taken to act on the will between enquiry and judgment, by distorting our ‘measures of probability’ themselves. Locke’s approach is more common-sensical than that of Descartes, but the psychology of motivated error is a hard nut which he also failed to crack.