Liberty,necessity and the foundations of Hume’s ‘science of man’

Introduction

In this article I put forward an interpretation of Hume’s section ‘On Liberty and Necessity’ in his Enquiry Concerning Human Understanding. I suggest that it can be read not only as a contribution to the philosophical debate about free will and determinism, but also as addressing the foundational issues of a characteristic 18^th-century enterprise,namely the ‘science of man’. More specifically, it can be read as a summary of Hume’s attempt to place this science on an experimental footing, with an awareness of the epistemological lessons he has drawn in the previous sections of the Enquiry. This interpretation fits with an overall reading of the Enquiry as responding to the epistemological problems arising in the context of contemporary ways of knowledge production. Some of the central problems of the Enquiry, like probability, induction and causation, have an obvious relevance in this context, but it may seem secondary in other contexts like those of miracles, providence, or sceptical philosophy. And it is not usual to read the section on liberty and necessity as responding to such problems either. By showing that it is fruitful to do so, I hope to take a step towards establishing the coherence of this work as read from this perspective.

Recent interpretations typically locate the relevance of this section in terms of its contribution to the free will debate, either in historical or in argumentative philosophical contexts. ¹ Although it has also been noted that what Hume has to say about liberty and necessity in Treatise of Human Nature is relevant for the experimental study of human nature, the discussion of those sections quickly slips into these conceptual frameworks (as, for example, in Pitson, 2006). I am not going to discuss this debate here, because I wish to offer an alternative framework of interpretation. The emphasis in the present case falls on Hume’s aspiration to provide a descriptive and explanatory theory of how human understanding works, and, if interpreted from this perspective, the questions of determinism, compatibilism and libertarianism do not arise for Hume, because taking a stance in this debate would commit him in questions of metaphysics that he would rather try to avoid. I am, therefore, in deep agreement with William Morris’ diagnosis:

Many of Hume’s readers miss what is so distinctive about his approach to philosophy because they are interested in metaphysical questions and take him to be interested in these questions in the way they are. In doing so, they fail to see that part of Hume’s attack on traditional approaches to philosophical questions consists in shifting the ground of discussion from what he regards as incoherent metaphysics to the only area where he believes we can have a fruitful discussion – where we have a clear understanding of the cognitive contents of the central ideas involved. (Morris, 2008: 473 f.)

Here I do not wish to quarrel with metaphysical reconstructions. I believe that it is possible to contextualize this section in a different way, and that by so doing a further layer of its significance becomes accessible. As I will argue, the section on liberty and necessity is about the epistemology of social cognition, i.e. about the extent to which, and the ways in which, knowledge of the causes of behaviour is possible. It is thus relevant for the methodology of a science of human nature due to its focus on the following two questions. What kind of knowledge can we hope for about human beings, and how should we pursue it? What are the meaningful questions that can be asked, and what is beyond the reach of this kind of inquiry? By answering these questions Hume sets the scope and limits of this science.

Questions of method in moral philosophy

There is considerable disagreement on the methodological ideal Hume observes in his inquiries. Newtonian natural philosophy and Baconian natural history may both seem equally plausible candidates, depending on which aspects of Hume’s methodologically relevant passages are emphasized. ² According to the 18^th-century classifications of knowledge, ³ Hume’s science of man belongs to the genre of moral philosophy that discusses phenomena characteristic of human beings qua moral beings – and not as natural or physical entities. This is the reason why he begins his investigations from what is universally accessible from a personal point of view, i.e. from perceptions. From the perspective of human beings as moral agents the natural philosophical and physiological explanations of sub-personal processes are inaccessible and of no explanatory use. Morality is a property of persons, thus in order to be meaningful its proper study requires explanations at the personal level in terms of what is accessible from the perspective of moral agents. This perspective ensures the autonomy of moral philosophy: it is an independent body of knowledge of phenomena that can be studied from different angles. ⁴

Following the experimental method of reasoning in moral subjects, as the Treatise’s subtitle suggests, obviously requires some experimental basis. This cannot be introspective, however, because as he notes reflecting on mental processes distorts them (Hume, Essay Concerning Human Understanding [hereafted cited as EHU], 2000: 1.13), so self-observation, let alone contrived inner experience, is no appropriate way of finding the principles of human nature that Hume wishes for. It ‘wou’d so disturb the operation of my natural principles, as must render it impossible to form any just conclusion from the phænomenon’ (Hume, Treatise of Human Nature [hereafter cited as THN], 2007: Introduction.10). Introspective phenomenological evidence – how things seem or feel to be – can play at most a secondary role among the experiential resources available while constructing a science of man on Humean grounds: what introspection tells us should be vindicated by the independent means of philosophical reasoning and should not be accepted at face value as its starting point. In order to avoid this, we need to rely on historiography as

… [i]ts chief use is only to discover the constant and universal principles of human nature, by showing men in all varieties of circumstances and situations, and furnishing us with materials from which we may form our observations and become acquainted with the regular springs of human action and behaviour. These records of wars, intrigues, factions, and revolutions, are so many collections of experiments, by which the politician or moral philosopher fixes the principles of his science, in the same manner as the physician or natural philosopher becomes acquainted with the nature of plants, minerals, and other external objects, by the experiments which he forms concerning them. Nor are the earth, water, and other elements, examined by Aristotle, and Hippocrates, more like to those which at present lie under our observation than the men described by Polybius and Tacitus are to those who now govern the world. (EHU, 2000: 8.7)

The role of historiography in moral philosophy is thus methodologically analogous with that of experiment in natural philosophy in an important way: both present their objects in various situations and furnish us with empirical material to theorize on. If our focus is on human nature, human history provides us with the variation of circumstances in which the characteristics of human nature can be identified and studied so as to establish the principles of its causal contribution in particular situations.

History establishes theoretical conclusions in very much the same way as experimental natural history, which entails that moral and natural philosophies can be built on similar methodological grounds. In the light of the above passage it is easy to conceive of historians as reporting experiments made on human nature – and while there may be no experimenter, there are reported events that can be treated methodologically as experimental findings. Therefore, ‘experiment’ in this context may well be metaphorical, but what really matters is that historical records are methodological equivalents of detailed experimental histories. Strictly methodologically, a moral philosopher can use historical works in very much the same way as a natural philosopher can rely on experimental histories produced by others. Newton, for example, selected from among Boyle’s and Hooke’s experimental findings to juxtapose them with available optical theories so as to gain new insights while working on his own theory of light. This is, as Kuhn says, ‘a non-Baconian use of Baconian experiment’ (Kuhn, 1977: 50): it proceeds not inductively but by contrasting empirical material with existing theories. This practice is quite consistent with the use of history Hume proposes in the study of human nature, both in its use of second-hand experience and in its juxtaposition of experience with existing theories. What thus becomes crucial in each case is to identify the relevant and reliable parts of histories. This poses a common problem, again, both to moral and to natural philosophies, one not to be discussed here, namely the role of testimony in cognition. ⁵

In order to proceed correctly in moral philosophy, ‘[w]e must … glean up our experiments in this science from a cautious observation of human life, and take them as they appear in the common course of the world, by men’s behaviour in company, in affairs, and in their pleasures’ (THN, 2007: Introduction.10). Historical records and direct observations can both provide the relevant material, but the former surpass the philosophical importance of observing particular cases, because history offers richer, and thus potentially more conclusive, empirical material for philosophical reasoning. Reporting large-scale transformations historiography presents several cases of a cause followed by an effect, and thus it provides a much broader and more effective basis for inductive generalizations than everyday experience or the observation of particular instances. The latter ones are much more likely to deviate from general regularities due to the influence of idiosyncratic circumstances (see Hume, ‘Of the Rise and Progress of the Arts and Sciences’ [hereafter cited as RPAS], 1985: 111 ff.), and they are thus less reliable sources for inferring the principles of human nature.

Experiments thus ‘judiciously collected’, philosophical work goes on with comparing them (THN, 2007: Introduction.10). ‘[T]he observation of several parallel instances’ (EHU, 2000: 8.13) and finding analogies between them gives the chance of reducing phenomena to the underlying principles productive of them (ibid.: 4.12). This enterprise makes sense only against the background presupposition of the structural uniformity of human nature. This commitment arises from Hume’s methodological rules 4 and 5 in the Treatise (THN, 2007: 1.3.15.6 f.), which claim that the same effect must be traced back to some similarity in their causes, a conviction that is extended to the study of human nature, too: ‘human nature remains still the same, in its principles and operations. The same motives always produce the same actions: The same events follow from the same causes’ (EHU, 2000: 8.7). This commitment is a methodological sine qua non for exploring the natural historical category of a moral being, but it should not be seen as an a priori category of a universal human nature in any substantive sense of some human essence. ⁶ What Hume’s project presupposes is not that human beings are the same sans phrase, only that there is a descriptive, natural historical category, identified on the basis of the similarities among its members that are constituted by qualitatively similar ingredients accessible by comparative analysis. And this is what specifies the cognitive benefit to be expected from Hume’s science of man: the ‘delineation of the distinct parts and powers of the mind’ (ibid.: 1.13).

Hume clarifies how to use the experimental basis in analogical reasoning so as to arrive at the principles of human nature, and then at the explanation of moral phenomena. The method here is a kind of analysis and synthesis:

By means of this guide [i.e. historical and everyday observations of human behaviour], we mount up to the knowledge of men’s inclinations and motives, from their actions, expressions, and even gestures; and again descend to the interpretation of their actions from our knowledge of their motives and inclinations. The general observations treasured up by a course of experience, give us the clue of human nature, and teach us to unravel all its intricacies. (EHU, 2000: 8.9)

We can never be allowed to mount up from the universe, the effect, to Jupiter, the cause; and then descend downwards, to infer any new effect from that cause … The knowledge of the cause being derived solely from the effect, they must be exactly adjusted to each other; and the one can never refer to anything farther, or be the foundation of any new inference and conclusion. (EHU, 2000: 11.14)

Analysing causes from effects cannot proceed arbitrarily: particularly, we cannot analyse phenomena into the characteristics of some deity – a point to which I will return below. What we can do, is to collect relevant phenomena, find analogies between them and ascribe those analogies to similar causes, thereby reducing a variety of phenomena to regular principles underlying them. But our knowledge cannot transcend what we can infer on an analogical basis from the effects themselves.

Overlooking the centrality of this ascending and then descending method in Hume can lead to the unjust allegation that there is a general instability in his thought resulting from the unclear relation between the science of man and history: the principles of the former sometimes seem to follow from historical observations, sometimes they serve as the explanation of historical events. ⁷ This is, however, not due to any intrinsic instability in Hume’s thought, but results from the methodological status of the principles themselves. We gain them from phenomena in the phase of analysis by comparative means, but in the phase of synthesis we use them for the purposes of explanation. The outlook in the two phases is opposite: in the first it turns from phenomena to principles, in the second from principles to phenomena. The epistemic aims are thus different: by analysis we aim at lawlike principles, by synthesis we aim at the explanation of phenomena by deriving them from these principles.

The methodological core idea is now visible. In some way salient human phenomena, explananda, are collected from history and observation, and then compared; if analogies and similarities are found, they are ascribed to some principles of human nature that are also compared, grouped and resolved into more general ones. These structurally fundamental principles of human nature are the proper aim of inquiry in the science of man. Methodologically speaking this is a commitment to processing empirical material, in an analogical way, on the assumption of structural uniformity, with an attention to the causal contribution of structural elements. Or in other words: Hume’s enterprise consists in identifying the functional ingredients of human nature and their characteristic role in producing human action and internal functioning. ⁸ This inquiry reveals the principles of human nature to be used as explanans in moral philosophy, and to explain thereby why our impressions and ideas follow one another in the order they do.

Hume’s principles of human nature are, then, qualitatively different: they are identified by their distinctive contribution to the chain of ideas and impressions. Due to the qualitative orientation of Hume’s method of analysis–synthesis, it is linked to the method Newton advertises in query 31 of the Opticks: ⁹

By this way of analysis we may proceed from compounds to ingredients, and from motions to the forces producing them; and in general, from effects to their causes, and from particular causes to more general ones, till the argument end in the most general. This is the method of analysis, and the synthesis consists in assuming the causes discovered, and established as principles, and by them explaining the phenomena proceeding from them, and proving the explanations. (Newton, 2004: 139)

As Newton himself here points out, he uses this method in the first two books of Opticks ‘to discover and prove the original differences of rays of light’, that is, to discover qualitative differences. This is what the method of analysis–synthesis, so understood, can reveal: causes belonging to different kinds. And while Newton applies this method in accounting for natural phenomena, he has no doubt that by doing so ‘the bounds of moral philosophy will be also enlarged’ (2004: 140). This method is analogical; it proceeds by the observation and comparison of different rays of light with respect to various properties like ‘refrangibility, reflexibility, and colour, and their alternate fits of easy reflexion and easy transmission’ (ibid.: 139). ¹⁰ Comparing rays of light reveals their determinable properties which, once determined, can be used in constructing explanations. This is also the central tenet of Hume’s method. ¹¹ Comparing moral phenomena results in determinable properties of human nature like the faculties of imagination, sympathy, moral sense, etc.; determining how these properties are actually instantiated in different circumstances provides the explanatory principles for the particular phenomena themselves.

This method can establish a link between natural and moral philosophy, and establishes them as methodologically continuous fields of knowledge in which the nature and extent of knowledge are the same. ¹² What I am suggesting here, however, is not that Hume’s method is explicitly modelled on Newton’s Opticks. I would rather say that it fits well with post-Newtonian methods of natural inquiry that became dominant in Scotland in the first half of the 18^th century (see Wilson, 2009: 80 ff. and 141 f.), and it is plausible to think about Hume’s method as being part of this strand of thought. For example, Hume’s method is quite similar to the methodological credo followed by William Cullen, a prominent member of the Scottish intellectual establishment and Hume’s friend and physician, in his influential chemical investigations. Cullen distinguishes between two kinds of analysis. ¹³ Analysis into integrant parts, i.e. a quantitative analysis into mereological proper parts, is the business of the mechanical philosopher, who studies substances as aggregates composed of homogeneous ingredients. Analysis into constituent parts, however, studies substances as mixtures, as being composed of qualitatively different components whose combination results in qualitatively different substances. The proper domain of this latter kind of analysis is thought to be the range of phenomena in which the mechanical approach cannot yield an explanation. And this is the kind of analysis that gives the methodological key both to 18^th-century ‘philosophical chemistry’ in Scotland, and to Hume’s method of studying human nature. ¹⁴

Three perspectives on human action

The continuity of moral and natural philosophy is founded on the fact that the phenomena they study are part of the same causal order. In explanations these phenomena are referred to as equal members of the same causal chain, as Hume’s famous example of a prisoner shows: his hopes for freedom are equally frustrated by the physical properties of the bars and the determination of the guard (EHU, 2000: 8.19) – natural and moral properties concur in making the punishment inevitable. Our reasoning about moral and natural phenomena is thus continuous. Inquiry in both fields of study is based on the idea of a necessary connection that arises from the impression we acquire due to experiencing constant conjunctions between phenomena. Our natural causal reasoning is based on this necessity, and this is the foundation of all theoretical causal cognition concerning the moral and the natural world (ibid.: 8.5). Among the phenomena studied by moral philosophy one can find constant conjunctions and there are also exceptions to the observed regularities – just like in natural philosophy. There is thus no reason to believe that the two fields of study should be methodologically different (ibid.: 8.19; 8.35). Character traits (virtues or vices) and motivations as causes are regularly linked to behaviour as effect, and these lawlike regularities are of the same kind as those on which reasoning about natural phenomena is founded, and they support the same kind of reasoning. The limits of human cognition are thus not narrower concerning moral than natural phenomena and the certainty of our knowledge about them does not differ either.

A third-person, observer perspective, from which phenomena are described and investigated, suggests the same necessity in both kinds of phenomena. What needs to be emphasized is: it is the same kind, but a very specially understood ‘necessity’. Hume’s concept of necessity differs from both the intuitive and the then-common understanding (see Ott, 2009), which is committed to a realist view that causation is a relation between events and has not much to do with our psychological propensities. Montesquieu (1989: 3), for example, defines the general concept of a law as a necessary relation that derives from the nature of things, but for Hume ‘[w]e do not understand our meaning in talking so’ (THN, 2007: 1.3.14.27). ¹⁵ For him necessity does not arise from the nature of things, or more precisely: given our cognitive capacities, the idea of necessity we can have is not that of something obtaining independently of the observer in the things themselves. ¹⁶ We do not have cognitive access to ‘natures’, only to the phenomenal properties of things, and we are sensitive to the regularities in their behaviour (EHU, 2000: 8.21) – we can get used to them. Given Hume’s epistemological commitments, the idea of necessity can only be traced back to the impression arising due to the repeated experience of constant conjunctions in the world surrounding us (ibid.: 7.28). Regarding human actions we acquire this idea from the insight that actions

… have a regular conjunction with motives and circumstances and characters, and as we always draw inferences from one to the other, we must be obliged to acknowledge in words, that necessity, which we have already avowed, in every deliberation of our lives, and in every step of our conduct and behaviour. (EHU, 2000: 8.22)

Hume here transmutes a question traditionally belonging to the realm of transcendent metaphysics into an empirical one – a move he quite frequently makes. He explains the idea of necessity we can acquire at all, given human cognitive capacities. As Howard Stein (2000: 261 f., 269 f., 277) shows, Newton does something similar: traditional metaphysical questions, like, for example, those concerning space or God, are turned into empirical ones in his hands, and they are thus transferred from the field of metaphysics to natural philosophy. ¹⁷ His strategy is to interpret observations as evidences grounding certain probable inferences concerning these questions. And Hume treats the foundational concepts of natural philosophy (e.g. causation, force, solidity, etc.) in a like manner: the genealogy and content of these concepts are not explored as a priori metaphysical issues, but as questions pertaining to an experimental science of man. ¹⁸

The concept of necessity is also subjected to this revision: Hume connects it to the observer’s point of view. From this perspective human beings are sensitive to regularities as one of our basic abilities is recognizing resemblances and getting used to them. For Hume this sensitivity explains our reliance on inductive inferences and causal beliefs, but this is also the foundation of conventions, institutions and the legitimacy of political power. Moral philosophy as an explanatory project is also based on the observer’s point of view, and it provides the means for understanding everyday social practices, and also our second-person, participant perspective which informs our taking part in them. From the observer’s point of view we detect and describe human actions and by comparing them we arrive at a conclusion concerning underlying principles that explain them and guide our reactions and expectations in the social world. The principles are founded on past experience which teaches us that the world of human action and interaction, just like the natural world, is full of detectable regularities, and participation in social interaction tacitly relies on them (EHU, 2000: 8.17). If there was no necessity, in this special sense, for us to rely on in these matters, then it would be ‘almost impossible … to engage, either in science or action of any kind’ (ibid.: 8.18) because ‘this experimental inference and reasoning concerning the actions of others enters so much into human life, that no man, while awake, is ever a moment without employing it’ (ibid.: 8.17).

Considered from the present perspective, Hume’s remarks that disputes on liberty and necessity are purely verbal (EHU, 2000: 8.1 ff.) can be understood in the context of the 18^th-century ideal of public science. There is evidence that Hume himself adhered to this ideal; at least he played an active role in the Philosophical Society of Edinburgh, which was organized according to the principles of this ideal; he even prepared some of its volumes for publication. ¹⁹ According to this ideal the proper epistemic foundation of inquiry is not scholastic authority as fossilized in canonical books, but experience accessible in public spaces, in the form, say, of publicly performed experiments. Quarrelling over words is only for making men enemies; therefore, instead of relying on the authority of books, one should read directly the Book of Nature itself (see, for example, Stewart, 1992: xxi; Shapin, 1996: 69). And this is what Hume suggests in this section of the Enquiry: instead of immersing ourselves in abstract reasoning and scholastic debates over words, we should have tenable definitions, based on a careful consideration of what we experience and how we act. This is the way to resolve the purely verbal conflict between liberty and necessity, i.e. on the basis of common and publicly accessible experience, from the observer’s point of view.

Now, experience teaches that it is only regularity on which our idea of necessity is founded – in the natural world as well as in human matters. On this basis, however, it is problematic to proclaim Hume a kind of determinist. ²⁰ Hume’s questions do not concern whether will or human action as such is free or determined, rather he asks: what kind of knowledge is possible on human actions and how is it possible? This question does not belong to transcendent metaphysics, as it does not ask whether the will itself is free: this is not a relevant issue for Hume. He concurs with the widely held view that liberty means ‘a power of acting or not acting, according to the determinations of the will’ (EHU, 2000: 8.23) – whether the will itself is free is not the question that bothers Hume, but the connection between action and motivation. Therefore the problem is epistemological and pertains to two fields of knowledge. First, philosophical knowledge – i.e. the knowledge of causes – presupposes the possibility of inference from causes to effects and vice versa. Causal inference is possible only if there are necessary connections grounded in the recognition of constant conjunctions as observed in history and in everyday life. Second, when not immersed in philosophical investigation but navigating the social world we rely on a similarly empirically grounded but implicit and practical knowledge and inferences. Our actions are unknowingly governed by the principles that philosophical investigations can make explicit. Philosophical knowledge and everyday interactions have the same foundation: the recognition of regularities and constant conjunctions in human behaviour.

Without ‘[n]ecessity, according to the sense, in which it is taken here’ no causal knowledge is available for human beings – this is the only way we can draw causal inferences, and Hume thinks that its practical and cognitive relevance has not been and cannot be denied (EHU, 2000: 8.21). Yet, there is a tension if the image of human actions one can get from the observer and the participant perspectives is compared with what we get from a first-person introspective perspective on our own behaviour and internal functioning. When contemplating from an observer perspective, repeated experience gives rise to an expectation that urges us to infer the effect upon experiencing the cause (ibid.: 7.30). But when doing something, we do not feel the connection between our motivations and actions inevitable (ibid.: 8.21). We feel instead something else, namely some ‘looseness or indifference’ whether to perform an action or not, and thus ‘we imagine we feel that the will itself is subject to nothing’. This is how ‘we feel a liberty in ourselves’ – yet, from the observer and participant perspectives we can fairly securely infer what action to expect given one’s motivations (ibid.: 8.22 n. 18). Therefore, inquiries into moral matters should not begin with the appearances of the first-person perspective, not from hypotheses ungrounded in analysis, but from regularities as found in repeated experience (ibid.: 8.22; see also Hume, Enquiry Concerning the Principles of Morals [hereafter cited as EPM], 1998: 1.10). On this basis inquiry reveals that liberty in the sense of indifference is not a fundamental feature of human nature but ‘a false sensation or seeming experience’ (EHU, 2000: 8.22 n. 18). This view is essentially the same in the Treatise: in the definition he gives, will is proclaimed to be not a faculty, but an impression that arises when we ‘knowingly give rise’ (THN, 2007: 2.3.1.2) to some motion or perception. ²¹ This feeling does not arise while contemplating actions, only while performing them, and this is why we imagine that in case of voluntary actions liberty as indifference prevails (ibid.: 2.3.2.2) – i.e. that will is not part of a chain of causes and effects. So understood, talking about the freedom of the will comes close to a category mistake: a faculty could be free in the sense of standing outside the chain of causes and effects, but it makes no sense to say that an impression is free.

This impression of indifference has no cognitive relevance for two reasons (EHU, 2000: 8.24). First, the idea we derive from this impression is inconsistent with the facts that are accessible from external observation and the common course of action. Introspection cannot be a good guide to how things stand, if one is in the business of charting our ‘mental geography’ (ibid.: 1.13). Although phenomenology, i.e. how things seem or feel from the first-person point of view, may seem to play a central role in Hume’s theory, generally, as we have seen above, it is not a reliable guide in exploring the principles of human nature (see ibid.: 1.13, and THN, 2007: Introduction.10). Phenomenology is not always reliable in distinguishing ideas and impressions either. Hume himself says that ideas can be as vivacious as impressions, and impressions can sometimes be mistaken for ideas (e.g. ibid.: 3.1.2.1), most significantly perhaps in the case of calm passions (e.g. ibid.: 2.3.4.8) that play a central role in moral cognition. We are similarly, but more profoundly, mistaken about the will.

Second, the idea is not consistent with itself either (EHU, 2000: 8.24 f.). The idea of liberty as indifference denies necessary, i.e. regular and constant, connections between motivation and action. If we identify liberty with standing outside the chain of causes and effects, then it proves to be ‘the same thing as chance’ (ibid.: 8.25). But actions independent of motivations could not be free but random, and random behaviour cannot be free, because, as is well known, ‘mad-men have no liberty’ (THN, 2007: 2.3.1.13). And this is why the impression of liberty as indifference is a false sensation: only a confused idea can arise from it, which is useless in the contexts of reasoning and action.

With his commitment to the authenticity of the observer and participant perspectives as opposed to the first-person perspective, Hume endorses the primacy of a scientific image of human nature as opposed to its manifest image. ²² What we can access introspectively is not validated by careful observation and analysis, and it is not reflected in our interpersonal interactions. Note, however, that this argument can only be conclusive if one preliminarily admits that the proper method of inquiring into moral matters is experimental in Hume’s sense. With this kept in mind the philosopher’s task is to begin with observable regularities, and it continues with tracing them back to their causal sources, and ends by explaining phenomena with the principles thus reached. We cannot begin elsewhere, and, particularly, we cannot presuppose the truth of the manifest image, because it is inconsistent with plain fact and itself. No inquiry can proceed on this presupposition which contradicts our natural commitments made explicit in our practices and which is a confused idea in itself.

Theodicy and the practice of moral evaluation

Ignoring the false sensation of liberty as indifference does not threaten liberty in the sense in which it ‘concerns us to preserve’, namely liberty as opposed to ‘force, and violence, and constraint’ (THN, 2007: 2.3.2.1, EHU, 2000: 8.25). This is the kind of liberty we need for the possibility of moral evaluation, and external constraints deprive us of this, because ‘actions are objects of our moral sentiment, so far only as they are indications of the internal character, passions, and affections; it is impossible that they can give rise either to praise or blame, where they proceed not from these principles, but are derived altogether from external violence’ (ibid.: 8.31). And, therefore, Humean moral evaluations also presuppose Humean necessity, because it is possible to infer causally from an action to its internal motivation only if constant conjunction underpins the inference. Moral evaluation of an action is possible only if we deny liberty as indifference and affirm independence of external constraint.

In the second part of section VIII of the Enquiry Hume does two things. First, he illustrates that our moral practice is based on the experience of constant conjunction and on analysis–synthesis, and thus it is continuous with philosophical inquiry. Moral evaluation springs from the action’s motivation, which is inferred on the basis of previously observed conjunctions between actions and motivations. Were there no such conjunctions, actions could not provide evidence for motivation, and consequently for moral evaluation (EHU, 2000: 8.30). Inferences from action to motivation presuppose a constant conjunction between types of action (virtuous or vicious) and relatively stable types of motivation (virtues, character traits, passions), a conjunction that gives rise to the impression of necessity in moral phenomena. Thus Hume shows that our moral practice is founded on the same idea of necessity as our reasoning about moral (and natural) matters.

Second, the lesson he offers about moral practice leads to the problem of theodicy and thus to a possible objection to Hume’s theory of necessity – at least in moral matters. If, relying on Humean necessity, actions are traced back to their motivations, and motivations to the circumstances influencing them, and the circumstances to other preceding circumstances, then we eventually arrive at ‘the Creator of the world’ (EHU, 2000: 8.32). From this angle two things can be said: either our actions cannot be wrong because ultimately they spring from a good cause, or they can be wrong, but then the Creator must share in the blame. As neither horn of this dilemma is possible, Hume’s doctrine of necessity is untenable – so argues Hume’s imagined opponent (ibid.: 8.32 f.).

Hume rejects the dilemma altogether. On the one hand he argues that the perspective of moral evaluation is much narrower in practice than to be sensitive to the entirety of this alleged causal chain. It is possible to argue that some particular wrong could arise from general laws that are otherwise right if considered from the perspective of the whole; or that some morally wrong action fits into the causal texture of the world so that it prevents something even worse or facilitates some future good: yet, these arguments cannot alter our natural moral sentiments that respond to particular actions. Moral judgements are founded on ‘the natural sentiments of the human mind’ (EHU, 2000: 8.35), and our moral sense, which responds naturally to character traits in relation to their contribution to sociability, is not susceptible to the conclusions of abstract reasoning.

On the other hand, prima facie it seems as if Hume could not find an effective argument against the second horn of the dilemma. Well, says Hume, if the Deity is the distant cause of our actions then it is unavoidable that the Deity be the cause of our vicious actions too. But the point of real importance comes only after this:

These are mysteries, which mere natural and unassisted reason is very unfit to handle; and whatever system she embraces, she must find herself involved in inextricable difficulties, and even contradictions. … Happy, if she be thence sensible of her temerity, when she pries into these sublime mysteries; and leaving a scene so full of obscurities and perplexities, return, with suitable modesty, to her true and proper province, the examination of common life; where she will find difficulties enow to employ her enquiries, without launching into so boundless an ocean of doubt, uncertainty, and contradiction! (EHU, 2000: 8.36)

Thus Hume’s response to the second horn is that reason is simply unequipped for solving the problem of theodicy, and thus the problem is placed outside the appropriate realm of philosophy, at least as it is defined by Hume’s standards of experimental reasoning. And this is the note on which this section ends: on the basis of our interpersonal practices and philosophical reasoning it is impossible to answer the questions about the Deity, and the Deity’s relation to the world. That is to say, these problems are unsolvable from a human point of view at all, because the solution, if there is any, lies outside the boundaries of reason.

Thus Hume commits himself concerning a question of central importance in contemporary natural philosophy, but in this chapter he approaches the problem from moral philosophy. This question pertains to the possibility of natural theology. Much of contemporary natural philosophy was driven by the conviction that God is the author of two books, namely the Bible and the Book of Nature (see, for example, Shapin, 1996: 104 f., 138 f.). Natural philosophers thought about themselves as studying the latter work through which God himself and his intentions can be revealed. ²³ While discussing theodicy, Hume considers the possibility of this kind of inquiry based on the study of human nature: to what extent, if at all, can God and his intentions be known from human actions? The question whether this kind of knowledge is possible is analogous with the question about the possibility of deriving a natural theology from our knowledge of nature – but the answers to the two questions are logically independent. ²⁴

Andrew Cunningham sees the role of natural theology in contemporary natural inquiry as so central that on this basis he denies the continuity of natural philosophy and modern science. Natural philosophy is about God even when its practitioners are not talking about him, a feature entirely uncharacteristic of modern science: ‘no-one ever undertook the practice of natural philosophy without having God in mind, and knowing that the study of God and God’s creation – in a way different from that pursued by theology – was the point of the whole exercise’ (Cunningham, 1991: 388). But while Cunningham is right about the central role of natural theology in natural philosophy, the relevance of this conclusion can be seen differently. Cunningham does not distinguish between the practice of natural philosophy and the ideology of this practice, and thus he identifies the practice with the meaning assigned to it. The practice of early modern natural philosophy can be rightly considered continuous with that of modern science; but the ideology attached to this practice, the way this practice is made sense of, and is ascribed significance, has been changed radically indeed. ²⁵

And this is the distinction with which one can bring to light Hume’s central point in the second part of this section. By declaring the problem of theodicy unsolvable and the science of man unfit for the contexts of transcendence he articulates an ideology of natural and moral knowledge: he detaches the questions of knowledge of nature and human nature from the questions of knowledge and knowability of God and his purposes. ²⁶ For Hume the problem of theodicy is relevant in this cognitive context, and it does not have a general import for the critique of religion here. ²⁷ Rather it drives toward a secular ideology of knowledge that distinguishes the modern natural and social sciences from the main tenet of early modern natural and moral philosophy.

Conclusion

In this article I have offered an interpretation that sheds a light on section VIII of Hume’s Enquiry in the context of contemporary ideals of knowledge. This reading offers three central considerations. (1) According to Hume’s ideal, a science of human nature is based on everyday and historical observations that are theoretically processed with the method of analysis and synthesis. Hume sketches this method as a qualitative analysis of human phenomena into their causal sources, which can be related to the methodological ideals of Scottish ‘philosophical chemistry’. Moral philosophy, as Hume sees it, is an enterprise that is continuous with natural philosophy both in methodology and cognitive aims. So it must be consistent with the common course of action – just as natural philosophy must answer to the course of nature. (2) Hume sketches three possible perspectives on moral phenomena. He considers the participant’s and the observer’s perspectives as continuous: if it is reached by the proper philosophical method, theoretical knowledge about human phenomena is congruent with our implicit practical knowledge about human matters as it is exhibited in everyday interaction. At the same time, he denies the adequacy of an introspective perspective, which leads us to conceptually and factually defective ideas about moral phenomena. (3) Hume argues that moral philosophy is an entirely mundane matter lacking any transcendent relevance. There is no justification for inferences drawn from human phenomena that are supposedly leading us to knowledge about how things stand outside this world, and there is thus no problem of theodicy that can be meaningfully addressed. Moral evaluation concerns human affairs only and its workings and functions can be accounted for in a philosophical way, in terms of appropriate causal relations. This is the only knowledge available to us from a human point of view. Moral philosophy, just like natural philosophy, is a human enterprise that belongs exclusively to the realm of human experience.