An Essay on Material Necessity(1)

Barry Smith

From: P. Hanson and B. Hunter, eds., Return of the A Priori (Canadian Journal of Philosophy, Supplementary Volume 18), 1992.

At the one extreme stands Hume, who - at least according to common conceptions - rules out entirely the possibility of material or non-logical necessity, and who therefore rules out also the possibility that we might enjoy that sort of certain knowledge that earlier philosophers had assumed as a matter of course to be correlated therewith.(2) At the other extreme stands Adolf Reinach, the hero of our present story, who defends the existence of a wide class of material necessities falling within the domain of what can be a priori known. More precisely, Reinach holds that there are certain categories of entity whose factual instantiation brings with it as a matter of necessity the instantiation of certain other correlated categories. The instantiation of the category colour necessitates in this fashion the instantiation of the category visual extension. Such categories are to be understood, from Reinach's point of view, after the fashion of Plato's intelligible forms. Thus they constitute an autonomous, natural order that is not capable of being invented or constructed through acts of man; they are such that actual things and events may serve to instantiate them more or less completely or perfectly; and they are such that we enjoy in relation to them (and therefore also in relation to their instances) a special sort of non-inductive knowledge. More consistently than Plato, however, Reinach acknowledges the existence of systems of relations through which the given categories are tied together into larger wholes. Such categories are then intrinsically intelligible precisely as nodes in this system of relations of the given sort, in the sense that anyone who has experience (including imaginative experience) of relevant instances of these categories is implicitly aware also of a corresponding family of relations. It is such intrinsic intelligibility of the basic structures and interrelations of shape and motion, colour and sound, for example, which makes it possible for us to learn language and to acquire those sorts of a posteriori knowledge on which empirical science is based.

Midway between these two extremes stands Searle. Like Hume, Searle is unwilling to accept special, uninventable categories whose instances would stand in necessary relations to each other. Like Reinach, however, Searle believes that there are categories of entity whose factual instantiation brings with it as a matter of necessity the factual instantiation of other correlated categories. The corresponding categories are for him not eternal and uninventable, however, but human creations, and the necessities in question are purely logical consequences of the `constitutive rules' by which the given categories are brought into being. As for Vico and Hobbes, so also for Searle, it is only in relation to what a human being has made that he or she is able to enjoy certain knowledge.

The essay which follows has two principal goals. On the one hand I shall summarize Reinach's thinking in connection with the issues of apriorism and material necessity. On the other hand I shall seek to demonstrate that Searle does not, after all, occupy a stable and acceptable half-way house between Hume and Reinach. Searle, too, if he is to be able to do justice to the establishment and to the workings of constitutive rules, must on pain of circularity embrace something like the Reinachian Platonism outlined above.

There is a commonsensical assumption to the effect, other things being equal, we can infer from: `John promised to do such-and-such' to: `John is obligated to do such-and-such'. Humeans (most philosophers, today) hold that the inference in question is invalid, that an `ought' can under no circumstances be derived from an `is'. Reinach holds that our commonsensical recognition of the validity of the given inference flows from our possession of synthetic a priori knowledge of a certain relation of necessitation between the categories in question. Searle, for his part, would have it that we have made the institution of promising and that as a result certain inferences are valid because they reflect the man-made rules of that institution.

It will turn out that speech act phenomena manifest in a particularly clear form the necessitation-structures which will be at the heart of our theory of the a priori. Thus I shall start with a brief excursus on the somewhat obscure history of philosophers' deliberations on such phenomena. This history has, I believe, some independent interest of its own, and will provide in addition some useful background for those interested in pursuing further the underpinnings of the Reinachian theory here defended.

On the Theory of Speech Acts

In the fourth chapter of his De interpretatione, Aristotle draws a distinction between two sorts of sentences. On the one hand, he says, there are `statement-making sentences'. On the other hand there are sentences, such as for example requests, `in which there is no truth or falsity' (17 a 1-5). The latter, Aristotle affirms, can be dismissed from logic, since consideration of them `belongs rather to the study of rhetoric or poetry.' Aristotle's attitude here, which remained authoritative throughout the Middle Ages, had fateful consequences. For his remarks ensured that the treatment of non-statement-making sentences came to be banished not merely from logic, but from the realm of science in general.

His remarks seem to have been explicitly challenged only towards the end of the 18th century in the work on `social acts' of Thomas Reid.(3) Reid's ideas in this connection remained without substantial influence, however, and it was not until the early years of the present century that the project of a theory of linguistic action began to take more definitive shape. More precisely, it is in a monograph on the foundations of the civil law, published in 1913 by our hero Reinach, that we find the first truly systematic theory of the phenomena of promising, questioning, requesting, commanding, accusing, enacting, etc., phenomena which Reinach, like Reid before him, collects together under the heading `social acts'.(4) Certainly there were philosophers other than Reid and Reinach who had considered the nature of promisings and other social acts. Such philosophers sought, however in silent tribute to the assumption of Aristotle that only statement-making sentences belong properly to the realm of logic to reassign such phenomena to the realm of statements or declarations. Thus Bolzano, for example, considers the act of questioning as a statement, to the effect that the questioner `desires to receive instruction about the object in question'.(5) The act of promising, similarly, is seen by Hume, Lipps and others as the expression of an act of will or as the declaration of an intention to act in the interests of the party in whose favour the declaration is made. There is an obvious inadequacy of all accounts of promising along these lines, however, namely that they throw no light at all on the problem of how an utterance of the given sort can give rise to claim and obligation. The bare intention to do something has, after all, no quasi-legal consequences of this sort, and it is difficult to see why things should be different in reflection of the fact that such an intention is brought to expression in language.

Certainly promising and communicating one's intention to do something do have much in common. Thus both belong to the category of what Reinach calls spontaneous acts, i.e. acts which involve a subject's bringing something about within his own psychic sphere as contrasted with passive or receptive experiences of, say, feeling a pain or hearing an explosion.(6) Both are, moreover, acts which involve as a matter of necessity a linguistic utterance or some other overt performance of a non-natural (in the sense of conventional or rule-governed) sort. This does not hold of other spontaneous acts such as judging or deciding, nor even of the acts of cursing or forgiving,(7) but it does hold of apologizing, commanding, accusing, entreating etc. Spontaneous acts can accordingly be divided into two classes, which we might call internal and external, respectively, according to whether the act's being brought to overt expression is a separable or inseparable moment of the relevant complex whole.(8)

Spontaneous acts are in every case intentional, which is to say that there is a (not necessarily existent) object or objects towards which they are directed. Spontaneous acts may in addition however be directed (in a different sense, now) to a subject or subjects, and most typically to a fellow human being. Amongst such subject-directed acts we may distinguish further between those which are self-directable and those which are non-self-directable (the latter Reinach also calls other-directed or "fremdpersonal"). The former are such that the subject toward whom they are directed may be identical with the subject of the act (as in cases of self-pity, self-hatred, etc.). The latter, in contrast, demand as that toward which they are directed a subject other than the one who acts.

Non-self-directable acts may be either internal (for example feelings of envy or resentment), or external (for example acts of baptism or benediction). Certain acts having the property of being both external and non-self-directable are in addition such that the relevant utterance must of necessity not only be directed toward but also addressed to and thereby registered or grasped by the subject in question. A command must as a matter of necessity be received and understood by the one to whom it is addressed (something which does not apply, for example, to an act of baptizing, forgiving or cursing). A command, that is to say,

is an action of the subject to which is essential not only its spontaneity and its intentionality, but also its being directed towards other subjects and its standing in need of being grasped by those subjects. What has been said of commands holds also for requesting, admonishing, questioning, informing, answering, and many other types of act. They are all social acts which are, in their execution, cast by him who executes them toward another subject that they may affix themselves in his mind [einem anderen zugeworfen, um sich in seine Seele einzuhaken]. (Reinach 1913, p. 707, cf. Eng. pp. 19f.)
What is important about an action of this kind is that it
is not divided into the self-sufficient execution of an act and an accidental statement [zufällige Konstatierung]; rather it constitutes an inner unity of deliberate execution and deliberate utterance. The experience is here impossible in the absence of the utterance. And the utterance for its part is not something that is added thereto as an incidental extra; rather it stands in the service of the social act and is necessary in order that this should fulfil its announcing function [kundgebende Funktion]. Certainly there exist also incidental statements relating to social acts: "I have just issued the command." But such statements then relate to the whole social act, with its external aspect (op. cit., p. 708, cf. Eng. p. 20).
Social acts, then, for Reinach (exactly as for Reid), involve
activities of mind which do not merely find in words their accidental, supplemental expression, but which execute themselves in the very act of speaking itself and of which it is characteristic that they announce themselves to another by means of this or some similar external appearance. (Op. cit., p. 728, cf. Eng. p. 36.)
The closeness to Austin and later speech act theorists here is unmistakable. A promise cannot be the merely incidental expression or intimation of an act of will or of an intention, because the act which underlies a promise is such that it is simply (and as a matter of necessity) not able to exist outside the compass of a whole of just this sort. And similarly there is no independent and self-contained mental experience which is somehow brought to expression, incidentally, in the issuing of a command. Hence, a fortiori, a social act of this sort cannot be a mere report or declaration of such an experience.

To say that there are experiences of such a nature that they exist only if they or their contents are simultaneously brought to expression might seem controversial the very existence of mutual dependence here implies that it is possible to separate experience and expression only abstractly (that it is in a certain sense illegitimate to speak of two distinct factors here at all). In some cases, however, the existence of an underlying experience is clear. Consider, trivially, the very experience of bringing to expression itself. Acts of warning, apologizing, promising, commanding, and so on, are likewise, Reinach holds, necessarily such that they cannot exist except in and through the very act by which they are brought to expression.

Actions of promising and commanding possess, then, as Reinach would have it, not merely an external dimension of utterance and execution, but also an internal dimension, through which they are tied into the domain of mental acts. The given phenomena are further such that they presuppose or are founded on appropriate mental states, for example of belief, and also on states of other kinds, for example the relational state of authority. In this way phenomena of the given sort are multi-categorial: their constituent relations of necessitation span regions of (mental, linguistic, legal and behavioural, factual and normative) reality and combine together into new structures elements derived from each.

We are far from having dealt with every aspect of Reinach's theory here. Thus we could have considered his treatment of conditional acts, of sham and defective and incomplete acts, of acts performed jointly and severally, of social acts performed in the name of some other party, of the ways in which such acts can be overridden and undermined, and so on.(9) One aspect of Reinach's work to which we shall return is the way in which his account of social acts is complemented by a theory of legal formations and of the ways in which the universal categories of promise, obligation, etc. are related to the contingent and pragmatically motivated higher-order social acts of the positive law. Reinach's most important contribution, however, at least for our present purposes, is to have drawn attention to the central role of necessitation relations in the sphere of social action and to have grasped the fact that our synthetic a priori knowledge is in large part precisely knowledge of structures held together by relations of the given sort. Necessitation relations had, it is true, been recognized earlier, above all by Husserl and other followers of Brentano in the sphere of psychology. It was Reinach, however, and his fellow Munich phenomenologists, who demonstrated the pervasiveness of the corresponding structures and who showed also how an absolutely general theory of a priori knowledge can be built up on this basis.

The Ontology of Necessitation Relations

It is an intriguing question whether Reinach's work exerted a direct or indirect influence on the development of speech act theory in Oxford in the '40s and '50s.(10) More important for us here, however, is the nature and content of Reinach's theory itself. The Anglo-Saxon treatment of speech act phenomena arose, familiarly, at a time when philosophy was conceived as a rather informal matter of `conceptual geography'. Reinach, in contrast, treats of speech acts and related phenomena ontologically. That is, he sees promises, claims, obligations, etc., as entities of certain special sorts, embrangled together in systematic ways with entities of certain other special sorts (for example with people, and with acts and states). Social acts are part of the very stock of the world, and not for example mere reflections of our conventional ways of speaking about other, more humdrum entities. This ontological treatment of necessitation in re will generate, as we shall see, a corresponding ontological theory of the a priori.

The world contains promisings, claims, commands, requests, marryings, and relations of authority, just as it contains instances of biological and logical species such as lion and tiger or judging and inferring. The species which people the world can be divided further into two sorts. On the one hand are independent species, whose instances do not require instantiations of other species in order to exist.(11) Lion might be taken as an example of an independent species in this sense. On the other hand are dependent species such as smile or dent, whose instances do not exist in and of themselves but only in association with instances of complementary species of other determinate sorts. Wherever instances of dependent species exist, there exist also relations of ontological necessitation. The inference from promise to obligation reflects an ontological necessitation of this sort, as also (in Descartes' eyes at least) does the inference from cogito to sum. Furthermore, where such positive necessitation holds, there obtain also certain correlated constraints or exclusions in a negative direction. Thus for example since colour depends for its existence upon extension, it follows that it is excluded as a matter of necessity that two distinct colours should occupy the same area or volume of space. The family of necessitation-relations extends, in this way, to fill out the entire sphere of what, traditionally, had been seen as the domain of a priori knowledge.

There is a rigid, autonomous order to the association or connection of dependent and independent entities. As Husserl puts it:

It is not a peculiarity of certain sorts of parts that they should only be parts in general, while it would remain quite indifferent what conglomerates with them, and into what sorts of contexts they are fitted. Rather there obtain firmly determined relations of necessity, ... determinate laws, which vary with the species of dependent contents and accordingly prescribe one sort of completion to one of them, another sort of completion to another. (1900/01, pp. 244f., cf. Eng. p. 454)
Judging is an example of a dependent species in Husserl's sense: a judging exists only as the judging of some specific subject (as a smile smiles only in a human face). Promising, too, is an example of a dependent species. Here, however, we see that the dependence is multifold: a promise requires that there be at least also the instantiations of species claim, obligation, utterance and registering act, reticulated together with language-using subjects in a single whole. As Austin himself recognized, every speech act is dependent on its surrounding circumstances, as it is dependent also on being heard (on uptake, or `Vernehmung' in Reinach's terminology).(12) Much more than Austin, however, Reinach was able to do systematic justice to the entire family of such necessitation relations within the framework of a single theory.

Such necessitation relations may be either one-sided or mutual. As an example of the latter, consider the relation of reciprocal dependence between the claim and obligation brought into being automatically with the execution of an act of promising. Or consider the relation between what Reinach calls the internal and external moments of the promising act. Both internal intending act and external utterance-phenomenon are, as we have seen, not able to exist outside the compass of a whole of just this sort. Intending act and utterance-phenomena are accordingly only superficially similar to intending acts or utterance-phenomena of the sort that are able to exist outside the framework of a promise.

The given features of the promise are marked also by a special a priori quality: they seem to be something of which we know not merely through experiment or observation or induction but rather immediately and (in principle) on even cursory acquaintance with phenomena of the relevant sort. They can be grasped without further ado, in the way that we grasp, for example, that red is darker than yellow, that blue is not a shape, that nothing can be simultaneously red and green all over, and so on. The categories in question (promise, colour, shape, etc.) seem in some way to be prior to any given factual realization. The corresponding necessary laws can be grasped as necessary even in the absence of factual instantiations.

As the six Platonic solids constrain factually existing processes of crystal-formation, so the categories distinguished by Reinach as basic to the sphere of social acts constrain factually existing institutions. They provide, as it were, a restricted range of natural and in a certain sense inevitable alternatives within the framework of which institutions must then be formed.(13) Through deliberately contrived institutional arrangements, thus through acts of lawgivers and others, such uninventable categories may become modified, in particular cultures and times, above all in the sense that the range of their particular realizations may become narrowed in specific ways. Thus for example it may very well be that in some given society a contract acquires validity only when it bears a certain official stamp. As Reinach saw, however, each such modification must be in keeping with the intrinsically intelligible category which serves as its starting point: hence the given conventional modifications are themselves constrained by certain necessary and intelligible laws. Thus there could be no culture or society in which the validity of contracts is in general made conditional on the parties' undergoing painful surgery, or on the finding of a proof for some hitherto unproved conjecture in mathematics. The absurdity of such modifications is grasped immediately by anyone with even a cursory familiarity with entities of the sorts involved. It is an a priori absurdity.

Interestingly, we have to deal here with entities existing in different ways in time. Obligations, claims and for example marital ties are (relational and non-relational) states of the corresponding objects. Their dependence consists in the fact that they cannot endure unless relevant bearers exist.(14) The spontaneous acts discussed above, on the other hand, are instantaneous events: their dependence consists in the fact that they cannot occur unless their bearers exist. Now, however, we see that dependence relations can obtain not merely between states or events and their bearers, but also forwards and backwards in time, between the states and events themselves, and even between states and events of different bearers. Thus if, as a matter of fact, certain actions are performed by a suitably authorized speaker under such and such conditions, then as a matter of necessity there begin to exist certain claims and obligations involving this speaker (or his principal) and certain other persons. These claims and obligations then exist for a certain time thereafter, and it may be that an end to their existence can be brought about by further social acts (for example by acts of waiving a claim), whose corresponding effects will then be no less necessary and automatic (and grasped a priori as such).

It is to repeat once more Reinach's point in reflection of our prior knowledge of the necessary relations which here obtain that we are able to derive the `ought' of obligation (such and such a person ought to do such and such) from the `is' which is expressed in a certain factual statement (to the effect that an utterance-phenomenon of a certain sort has occurred at a certain time and under such and such conditions).

The Logical Conception of Material Necessity

There is a deep-rooted temptation, now, to suppose that such a priori necessities are one and all a matter of the logical relations which hold among the corresponding concepts. An account along these lines does indeed enjoy considerable purchase in relation to necessary laws in the purely formal spheres of for example mathematics. Such an account has however failed repeatedly when the attempt has been made to extend it to the necessary laws obtaining in such material spheres as those of colours or tones or social acts.(15) To say of the truths of mathematics that they are `formal' is to say that they are topic-neutral, that they are not specific to any given qualitative or material sphere.(16) Such truths can at least in principle be applied unrestrictedly to all qualitative or material regions (as one can count, for example, everything from ring nebulae to canzonettas). When once the transition is made to laws whose content involves a restriction to what is qualitatively or materially determined, however, the logical account can no longer be successfully applied.

This follows from the ineliminable presence of materially specific concepts (or corresponding terms) in the formulations of these laws themselves. On the logical account, propositions or sentences seemingly expressing material necessities are seen as being characterized by the fact that they can in every case be exposed as mere analytic truths which are entirely empty of content. Such exposure is effected via a process of stripping out defined terms, replacing each such term with a definiens consisting of more primitive expressions. `All bachelors are unmarried' is revealed as analytic in this way, by being converted to `All unmarried men are unmarried', a substitution instance of the logical truth: `All A's which are B are B'.(17)

If, however, we wish to hold on to the view that all the necessary propositions of the theory of (say) social acts are analytic in this sense, then we shall have to insist that all such propositions can be formulated in terms of at most one single primitive non-logical concept. For suppose that even two such concepts were needed, say and , neither (by hypothesis) definable in terms of the other. Consider now the propositions expressing relations between and , even such trivial relations as the relation of non-identity. These cannot (again by hypothesis) be analytic. For propositions expressing non-identity of concepts (green is not yellow, colour is not sound) are clearly not substitution-instances of logical truths as they stand. But there are (again by hypothesis) no defined non-logical terms which could be eliminated in such a way as to reveal the corresponding statements as logically true. But nor (again by hypothesis) can they be merely factual. On the logical reading of necessity, however, no further alternative is available, which implies that the original assumption that there are two (or more) such concepts must be rejected.

When, however, we consider the wealth of independent concepts involved in the theory of social acts concepts of language, of mental acts of intending, willing, registering, of mental states of sincerity and good faith, of obligation, claim, authority, concepts of action and performance, etc. then it is clearly impossible that the entire family of such concepts might be capable of being reduced, by a process of definition, to at most one single non-logical primitive. The corresponding laws, therefore, or the bulk of them at least, are matters of non-logical necessity, and our corresponding knowledge of such laws is synthetic and a priori.

Institutional Concepts and Constitutive Rules

If we cannot understand the necessary laws which obtain in the sphere of social acts in purely logical terms, then how are we to understand them? Reinach, as we have seen, adopts a Platonistic answer to this question, an answer according to which there obtain necessary and intelligible relations between categories of certain sorts (for example promises, claims and obligations) as these are realized in the world. It is in reflection of our knowledge of such relations that the derivation of `ought' from `is' in the case of the promise is grasped as valid a priori.

Between this Platonic extreme and the Humean position which would deny the given derivation tout court, we have the conventionalism of Searle. For Searle, too, there are necessities obtaining in the sphere of social acts. For him, however, these necessities follow from the special character of those forms of behaviour which involve what he calls `constitutive rules', rules whose adoption at a certain time by a certain culture or society brings into being the corresponding categories or forms of behaviour.

There are, Searle tells us, two different kinds of rules or conventions:

Some rules regulate antecedently existing forms of behaviour. For example, the rules of polite table behaviour regulate eating, but eating exists independently of these rules. Some rules, on the other hand, do not merely regulate an antecedently existing activity called playing chess; they, as it were, create the possibility of or define that activity. The activity of playing chess is constituted by action in accordance with these rules. Chess has no existence apart from these rules. (1969a, p. 131)
The same can be said also, from Searle's point of view, of the institution of promising:
The institutions of marriage, money, and promising are like the institutions of baseball and chess in that they are systems of such constitutive rules or conventions. (loc. cit.)
Searle is appealing here to an opposition between `institutional' and `brute' facts, the former being distinguished in this way: that they presuppose deliberate constitutive arrangements of the given sort. Many forms of obligations, commitments, rights and responsibilities are, Searle holds, a matter of institutional facts in this sense. As for Reinach, so also for Searle, the oughtness of obligation follows as a matter of necessity from the isness which is the making of a promise. For Searle, however, this is a definitional matter, for the making of a promise is for him by definition a case of acting according to certain conventional rules and in these rules the notion of obligation is involved in the relevant sense.

There are on this view no special universal and uninventable categories which our everyday acts of promising might reflect and by which the very institution of promising might somehow be constrained and made intelligible. The illusion that there are such categories arises only as a result of the fact that, because certain parts of reality have themselves been shaped by rules of the constitutive sort, we are able to utilize corresponding concepts and to apply them to corresponding pieces of behaviour.

There are clearly, however, as Searle himself must recognize, certain constraints on the constitutive rules that human beings can adopt. Thus, for example, we cannot have a rule which enjoins walking through walls, or travelling back in time, or making 2 + 2 make 5. Institutions are in this sense constrained by brute facts and also by the requirement of logical consistency of the underlying rules. It seems that for Searle, however, they cannot be constrained by anything other than this; for what could this something else be? All conceivable constitutive rules which are factually realizable and logically coherent must from his perspective be at least ex ante of exactly equal status, however absurd or unnatural they might seem.

Reinach, in contrast, has a means of doing justice not merely to the necessary laws and necessary constraints governing the field of institutional facts, but also to our capacity to grasp immediately the absurdity of institutional arrangements which violate such constraints (as we can grasp immediately the a priori absurdity of eating a phoneme or weighing a number). He holds, as we have seen, that there is a family of uninventable and intrinsically intelligible categories which serve as the necessary basis for rule-formation in the institutional sphere. Thus there could not, according to Reinach, arise institutions which did not reflect the basic categories of, for example, utterance-phenomena, claims, obligations, and their interrelations.

For Reinach promise and obligation are elements in a complex natural hierarchy of universally instantiable categories. It is in reflection of this autonomous hierarchy of categories that the relevant concepts and the rules we follow in speaking and acting have in large part arisen. These categories and their instantiations may involve (e.g.) linguistic elements, but they are not contributed by language or convention. On the contrary, the very practices involved in formulating and adopting conventional rules presuppose universal categories of the given sorts, not least those universal categories which concern rules and conventions themselves and all that goes together therewith. Clearly, on pain of circularity, we cannot hold with Searle that such universal categories could themselves have been invented via rules and conventions.

Note that Reinach accepts that certain institutional conveniences may in the course of history come to be attached to universal categories of the given sorts as these are realized in particular societies. Thus he is willing to concede to Searle that even a world which manifests the given uninventable universal categories in the realm of social acts might still have room for purely conventional arrangements built up on these, reflecting constitutive rules of the sort Searle favours.(18) Cases of concepts which are "purely conventional" in this sense might be: endowment mortgage, lien bond, football team-manager, and so on. These (we may reasonably suppose) correspond to no special universal categories, but are rather imposed upon the world via constitutive rules in the way that Searle describes. The criterion of pure conventionality here, a criterion which Reinach, too, could readily accept, is the possibility of our introducing the concepts in question via non-circular definitions expressible in terms of concepts which are truly and unproblematically more basic. For Reinach, however, we must by these means eventually arrive at basic institutional concepts, which is to say: institutional concepts not capable of being further defined on the institutional level. Ownership, presumably, is a concept of this sort; others might be: rule, obligation, benefit, exchange, utterance, uptake, understanding, agreement, preference, sincerity, and so on. Similar basic institutional concepts are required also e.g. in the realm of games - concepts such as winning, losing, playing, breaking a rule, and so on. And clearly such basic institutional concepts, or the corresponding natural categories, must be involved also where constitutive rules are formulated and adopted in the realm of the positive law, concepts such as command, decision, authority, consent, acknowledgement, jurisdiction, and so on.(19) Consider, now, the truths holding of such basic institutional concepts. The question for Searle is: are such truths purely conventional in the sense defined above? Clearly not: for the very formulation and adoption of constitutive rules presupposes concepts of the given sort. Are they, then, merely analytic? This alternative, too, can be ruled out, by the argument given above on the inapplicability of the analytic conception to relationships of material necessity. Can we, then, suppose that all such concepts can be defined in non-circular ways in terms of non-institutional concepts? Not at all, for then all institutional concepts would turn out to be thus definable, an outcome which Searle quite rightly rules out.(20) The only alternative which remains, therefore, is for Searle to accept that the given truths express irreducible material necessities of the Reinachian sort, that is, that they express necessary relations between certain uninventable sui generis categories. That Searle has not faced the necessity of drawing this conclusion follows from the fact that he has always already presupposed a rule-positing society, without ever asking how this society and its rule-positing practices came about - not, clearly, by prior rules).

Not only Reinach, then, but also Searle must accept the notion of basic institutional concept. Moreover, both must accept also the irreducible institutional categories to which such concepts must correspond. Where they might still disagree is in relation to the question as to where the line is to be drawn between what we have called purely conventional concepts (concepts which can reasonably be held to have been introduced by definition) and basic institutional concepts (concepts reflecting irreducible categories for which non-circular definitions cannot be supplied). Promising, in particular, is taken by Searle to be a purely conventional concept, where Reinach insists that it is basic.(21) Such borderline disputes need not detain us long, however, so long as it is accepted that we are in possession of clear cases on either side of the disputed border. For then it will at least have been established that there is an intrinsic order of categories that is prior to any imposed order which we might seek to contrive e.g. by definitional fiat. Recall, in this connection, the reconstructive endeavours of Whitehead and Russell in Principia Mathematica. Number, in their framework is defined (though not before *100 of volume II) in terms of certain other, not intuitively more basic concepts, including propositional function and type. That such a definition can be constructed (albeit, notoriously, at a certain price in terms of other disputed assumptions in the system) does not, from the Reinachian perspective, imply that numbers are what they are reputed to be in the context of that system. Rather, it shows that we have a certain free play in constructing definitional systems adequate for example to our deductive purposes in mathematics.

A Theory of the A Priori

Let us summarize the results of our deliberations as far as the phenomenon of a priori knowledge is concerned. Our thesis is as follows: that a priori knowledge relates first and foremost to the relations (above all relations of necessitation) which obtain between intelligible categories of for example colour, sound or shape. Man, it is suggested, is born with an innate capacity to discriminate between instances of categories of the given sorts for this is not itself something that could have been learned and hand in hand with this innate capacity goes the ability to grasp the associated relations of necessitation. One incidental advantage of the view in question is that it yields a systematic accounting of all the various types of a priori laws (laws relating to one-sided and mutual necessitation, laws of exclusion and compatibility, and so on), where more familiar treatments of the a priori are constrained to conceive the latter as amounting to little more than a collection of ad hoc examples. Our view amounts, as already suggested, to an ontological (realist, reliabilist) theory of the a priori: we enjoy synthetic a priori knowledge of, say, the basic relations between colours, because there are corresponding intrinsically intelligible relational structures in the world. This view, once the common possession of all philosophers, who rightly saw the existence of such structures as providing the necessary and sufficient conditions for the existence of their discipline, is diametrically opposed not only to standard Humean conceptions (which simply deny the existence of necessity in re), but also to the Kantian view, which sees the a priori as a result of the imposition of structure by experiencing subjects. Kant was, to be sure, correct in recognizing the central importance of synthetic a priori knowledge not only to philosophy but also to science and to our everyday experience. He was wrong, however, not only in his estimation of the scope of such knowledge which he saw as being restricted, effectively, to not much more than arithmetic, geometry (Euclidean) and mechanics (Newtonian) but also, and disastrously, in his account of where such knowledge comes from.


Austin, J. L. 1962 How to do Things with Words, Oxford: Oxford University Press.

Burkhardt, A. 1986 Soziale Akte, Sprechakte und Textillokutionen. A. Reinachs Rechtsphilosophie und die moderne Linguistik. Tübingen: Niemeyer.

Burkhardt, A. 1987 "Verpflichtung und Verbindlichkeit. Ethische Aspekte in der Rechtsphilosophie Adolf Reinachs", in Mulligan (ed.), 155-74.

Burkhardt, A. (ed.) 1990 Speech Acts, Meanings and Intentions, Berlin/New York: de Gruyter.

Husserl, E. 1900/01 Logische Untersuchungen. as repr. in Husserliana vol. XVII (ed., E. Holenstein) and XIX/1, XIX/2 (ed., U. Panzer), The Hague: Nijhoff, 1975/1984.

Mulligan, K. 1987 "Promisings and Other Social Acts: Their Constituents and Structure", in Mulligan (ed.), 29-90.

Mulligan, K. (ed.) 1987 Speech Act and Sachverhalt. Reinach and the Foundations of Realist Phenomenology, The Hague: Nijhoff.

Paulson, S. L. 1987 "Demystifying Reinach's Legal Theory", in Mulligan (ed.), 133-54.

Reinach, A. 1911 "Zur Theorie des negativen Urteils", in A. Pfänder, ed., Münchener Philosophische Abhandlungen, Leipzig: Barth, 196-254, repr. in Reinach 1989, 95-140; Eng. trans. as "On the Theory of the Negative Judgment", in Smith (ed.), 1982, 315-77.

Reinach, A. 1911a "Kants Auffassung des Humeschen Problems", Zeitschrift für Philosophie und philosophische Kritik, 141, 176-209, repr. in Reinach 1989, 67-94; Eng. trans. by J. N. Mohanty as "Kant's Interpretation of Hume's Problem", in Southwestern Journal of Philosophy, 7, 1976, 161-88.

Reinach, A. 1913 "Die apriorischen Grundlagen des bürgerlichen Rechts", Jahrbuch für Philosophie und phänomenologische Forschung, I/2, 685-847, repr. in Reinach 1989, 141-278; Eng. trans. as "The Apriori Foundations of the Civil Law", by J. F. Crosby, Aletheia, 3, 1983, 1-142.

Reinach, A. 1989 Sämtliche Werke. Kritische Ausgabe mit Kommentar, 2 vols., K. Schuhmann and B. Smith (eds.), Munich/Hamden/Vienna: Philosophia.

Schuhmann, K. 1988 "Die Entwicklung der Sprechakttheorie in der Münchener Phänomenologie", Phänomenologische Forschungen, 21, 133-166.

Schuhmann, K. and Smith, B. 1985 "Questions: An Essay in Daubertian Phenomenology". In: Philosophy and Phenomenological Research, 39, 763-93.

Schuhmann, K. and Smith, B. 1987 "Adolf Reinach (1883-1917)", in Mulligan, (ed.), 1-27.

Schuhmann, K. and Smith, B. 1990 "Elements of Speech Act Theory in the Work of Thomas Reid", History of Philosophy Quarterly, 7, 47-66.

Schulthess, D. 1983 Philosophie et sens commun chez Thomas Reid (1710-1796), Berne.

Searle, J. R. 1969 Speech Acts. An Essay in the Philosophy of Language, Cambridge: Cambridge University Press.

Searle, J. R. 1969a "How to Derive `Ought' from `Is'", in W. D. Hudson, ed., The Is/Ought Question, London: Macmillan, 120-134.

Smith, B. (ed.) 1982 Parts and Moments. Studies in Logic and Formal Ontology. Munich: Philosophia.

Smith, B. 1981 "Logic, Form and Matter", Proceedings of the Aristotelian Society, Supplementary Volume 55, 47-63.

Smith, B. 1986 "Austrian Economics and Austrian Philosophy", in W. Grassl and B. Smith, (eds.), Austrian Economics. Historical and Philosophical Background, London and Sydney: Croom Helm, 1-36.

Smith, B. 1990 "Towards a History of Speech Act Theory", in Burkhardt, (ed.), 29-61.

Smith, B. 1990a "Aristotle, Menger, Mises: A Categorial Ontology for Economics", History of Political Economy, supplement to vol. 22, 263-288.


1.  My thanks are due to Johannes Brandl, Patricia Donohue and Philip Hanson for helpful comments on an earlier version of this paper.

2.  For a more adequate discussion of Hume's views on this matter see Reinach 1911a.

3.  Cf. Schulthess 1983, p. 304; Mulligan 1987, pp. 33f.; Schuhmann and Smith 1990.

4.  Reinach's theory was in part inspired by the work on logic and ontology of his teacher Edmund Husserl. It incorporated also criticisms of Husserl's thinking derived from the Brentanian philosopher of language Anton Marty and from Reinach's friend and fellow student in Munich Johannes Daubert. On the pre-First World War Munich tradition of speech act theory see Smith 1990 and Schuhmann and Smith 1985. An important role in the development of Reinach's thinking was played also by his background as a student of law: see Schuhmann and Smith 1987, pp. 10-13.

5.  Wissenschaftslehre (1837), vol. I, § 22.

6.  See Reinach 1913, p. 706, Eng. p. 18. The issue whether the distinction between spontaneous and passive acts is an exhaustive one is not important for our present purposes.

7.  We leave aside here those acts of forgiving which are prompted by a request for forgiveness.

8.  Reinach 1913, pp. 707f., Eng. p. 20, Mulligan 1987, Schuhmann 1988. On the terminology of `inseparable moments' see Smith, ed. 1982.

9.  Cf. Mulligan 1987, pp. 78ff., Smith 1990.

10.  This question is not at this stage able to be decided. We do know, however, that Austin's early interest in German (and Austrian) philosophy was not confined to his work of translating Frege. Moreover, it is known that a copy of Reinach's Gesammelte Schriften was possessed by Austin's colleague Gilbert Ryle. This survives, with annotations, in the library of Linacre College in Oxford.

11.  See Smith, ed. 1982 for more details of the theory of dependence and independence as this applies to species.

12.  See Austin 1962, pp. 22, 52 and compare Reinach 1911, p. 213, Eng. p. 373, n. 11, 1913, pp. 707, 796, 801, Eng. pp. 19, 94, 102.

13.  This universality manifests itself also in the fact that we naturally tend to understand alien social orders in terms of the same restricted class of basic notions (claim, obligation, etc.), just as we tend naturally to see slightly irregular triangles and squares in terms of the standard concepts of triangle, square, etc.

14.  Claim and obligation are in this sense comparable to the individual accidents of the tradition. They differ from standard Aristotelian examples, however, in that they can have a multiplicity of bearers; moreover, they are able to be passed on from one bearer to another (claims and obligations can be inherited; they may migrate from one substance to another). Here, therefore, we have to do not with an individual necessary dependence of one entity upon another, but rather with what might be called a generic dependence, whereby it is required only that the dependent entity exist in association with some entity or other from a certain restricted class. (See, again Smith, ed. 1982.)

15.  The argument which follows is developed at greater length in my 1986, pp. 15-18.

16.  See my 1981 for further consideration of the meaning of the terms `formal' and `material' in the present context.

17.  This is no mere incidental mark of analyticity, but a statement of what it is for a proposition to be analytic. There are, to be sure, competing accounts as to what `analytic' might mean, drawn for example from Wittgensteinian `logical grammar'. None of these accounts has however succeeded in establishing itself as a clear and natural alternative to the Fregean account adopted here.

18.  Cf. Reinach 1913, pp. 801f., Eng. p. 104, where the parallels between Reinach's notion of enactment and Searle's notion of constitutive rule are especially clear. See also Paulson 1987 and Burkhardt 1986.

19.  Note that nothing in what follows turns on the question as to whether we have provided even partially adequate lists of basic institutional concepts here: the lists provided in the text are intended to serve as illustrative examples only.

20.  1969, p. 56.

21.  Searle has indeed offered a definition of "promising" in terms of other, more basic concepts (see his 1969, pp. 57ff.), though it has to be said e.g. in relation to clause 7. of this definition that it is not at all clear that the mere fact that someone intends that his utterance shall place him under an obligation to do such and such is sufficient to bring it about that his utterance will indeed have this effect. In regard to clause 9., similarly, we must ask whether the semantical rules of a language really can be such that a given sentence is correctly and sincerely uttered if and only if this sentence brings it about that the speaker is brought under an obligation.