The a priori

I will start from the problem of the material a priori. That obviously relates back to the disagreement which arose between Schlick and Husserl concerning judgements (or assertions) which were synthetically or materially a priori: judgements ‘discovered’ by Husserl, disputed by Wittgenstein, and openly contested by Schlick. The debate on this theme brought out certain elements of importance.

A. The key points of the debate. It has been subsequently recognized that the origin of this dispute is to be found in the nominalist tendency evinced by Schlick. It was this nominalism which led the author of the Allgemeine Erkenntnislehre to criticize the phenomenological theses on the subject of essences and the intuition of essences. In effect, in Husserl, it is thanks to the intuition of essences that one can establish ideal laws concerning both the eidetic truths of logic and mathematics – which he labels as formal and analytical truths – and the eidetic truths obtained by the method of free variation in the imagination and concerning the intuitive contents of particular sensorial domains (e.g. sounds and colours) – which Husserl labels as materially and synthetically a priori. All these ‘phenomenological’ truths govern as much the imaginative possibilities of the intuition as the ontological possibilities of effective states of things; as such, they allow the construction of material ^{Footnote 3} regional ontologies, in addition to the formal ontology of pure logic. It is precisely on this point that the critical literature has broadly focused for, to defend the linguistic theory of the a priori that is proper to logical empiricism, Schlick contended that these material a priori are in fact analytical and formal assertions anchored in the logical grammar of the language in which they are formulated, or even in the rules which regulate the use of the terms concerned (Reference Schlick, Schlick and WaismannSchlick, 1969a, Reference Schlick, Schlick and Waismann1969b).

A second element has emerged from the debate: this relates to a considerable divergence of opinion on the very nature of the dispute. Not only are all commentators not in agreement on the judgements or assertions that should be considered as examples of material a priori among the phenomenological propositions cited by Husserl (Reference Miraglia and LanfrediniMiraglia, 2006), but – and which counts even more – these divergences hark back, at least to a certain degree, to a primary discordance: that is, the different ways of interpreting the terms used to characterize those assertions said to be materially a priori (of the type: ‘Each sound has an intensity and a quality’, or ‘A surface is not at the same time red and green’) in comparison with assertions which are themselves of logico-analytical nature (‘If it is raining and it is cold, then it is raining’, ‘All bachelors are bachelors’, ‘All bachelors are unmarried men’, ‘A depraved man deserves no respect’). In other words, the discordance would emanate principally from the lack of clarity, of consonance, even, of the manner in which the pairs of opposed terms should be understood: logical/non-logical, a priori/a posteriori, analytical/synthetic, formal/material. Certain commentators have in fact argued that, following the aspects of the thought of Husserl and Schlick that we are considering, it is possible to give different characterizations to these dichotomies, and, in the final instance, to conclude either in favour of or against the existence of a material a priori distinct from a formal a priori.^{Footnote 4}

The third point which arises from the debate is the following: Schlick and the other neo-positivists based their positions initially, under the influence of Wittgenstein, on a notion of logical truth and analyticity which was too narrow to account for the difference between tautological statements of the logico-formal kind in the strict sense (e.g. ‘If it is raining, then it is raining’ and ‘All bachelors are bachelors’, which can be reduced to the logical forms of the kind ‘If p, then p’, ‘All the A's are A's’) and statements sometimes called truths ‘by essential predication’, such as ‘All bachelors are unmarried men’ (Reference Quine and QuineQuine, 1977: 109 s., 128).^{Footnote 5} The very refined tradition which links Bolzano with Husserl has always been cognizant of these distinctions; on the other hand, the line of argument linking Wittgenstein to the logical positivists has reached only over time, and under the pressure of the semantic approach of Tarski and Quine, a description of analytical truth (inasmuch as a truth founded on language or on the rules of linguistic usage) which is capable of rendering account of truths by essential predication. This is what Carnap would attempt to do in the 1950s: he would maintain that for a language to be thoroughly established, it must integrate the syntactic-semantic rules which govern the grammatical formation and the logical transformation of its expressions with ad hoc ‘meaning postulates’. He assigned to these postulates the task of fixing equivalence relationships of partial or total intensionality (synonymy) so as to align truths by essential predication (‘All bachelors are unmarried men’) with the logico-formal truths (‘All bachelors are bachelors’, Reference Carnap and CarnapCarnap, 1956).

The fourth and final constituent of the debate relates to the status to be accorded to propositions which Husserl considered as synthetically (or materially) a priori. Thanks notably to Peter Reference Simons and SimonsSimons (1992: 371–376), we now know that to understand the theory of the material a priori, account must be taken of the way Husserl, in the manner of Bolzano, characterizes the class of analytical assertions, which he sub-divides into the following two sub-classes: (i) logically analytical truths, or those which are analytical truths stricto sensu (also called ‘identical’ or ‘tautological’ truths) and (ii) analytical truths in a broader sense. The first are propositions where all the concepts (or words) which do not belong to the logic vocabulary may vary without the truth value of the proposition being changed (salva veritate aut falsitate). For example, the propositions ‘It is raining or it is not raining’ (‘p or not p’), ‘If it is raining, then it is raining’ (‘If p then p’), ‘A happy bachelor is a bachelor’ (‘An A which is a B is an A’) are logically (or strictly) analytical truths for they remain true whatever terms we may substitute for ‘it is raining’, for ‘bachelor’, and for ‘happy’. The second category of truths, that is analytical truths in the broader sense, are on the other hand those where at least one of the ideas or notions which appear in their affirmation may vary without the truth value of the proposition being subsequently affected. Thus ‘A depraved man merits no respect’ is analytical in the broader sense for only one of the notions it contains (in this example, that of man) can be replaced by any other notion (for example, by that of angel or entity) without the proposition being transformed into a falsehood.^{Footnote 6}

This distinction does not totally coincide with that between logical truths and truths by essential predication of the Carnap and Quine tradition. Broadly speaking, and without going into the details associated with the problem of reference,^{Footnote 7} one may say that logical truths of the kind ‘A bachelor is a bachelor’ (‘A is A’), ‘A happy bachelor is a bachelor’ (‘An A which is a B is an A’), (‘Each object is a bachelor or a non-bachelor’ (‘Each object is B or non-B’) are also logically (or strictly) analytical truths in the Bolzano and Husserl sense. On the other hand, one cannot say that their analytical truths in the broader sense (for example, ‘A depraved man merits no respect’) are comparable to truths by essential predication, such as ‘Bachelors are unmarried men’. That is why Simon comments on the relationship between the two types of classification by observing that ‘neither Husserl nor Bolzano, any more than Kant or Leibniz before them, were able to resolve satisfactorily the problem of hidden or implicit analyticity’ (Reference Simons and SimonsSimons, 1992: 374, my emphasis). It relates to the problem of the analyticity of truths by essential predication, which cannot be compared either to assertions of the type ‘A bachelor is a bachelor’ or ‘A happy bachelor is a bachelor’ (which are analytical in the strict sense), or to assertions of the type ‘A depraved bachelor deserves no respect’ (which is analytical in the broad sense). According to Reference Simons and SimonsSimons (1992: 374) it is precisely this failure which would have been at the origin of Quine's critique of analyticity.

Beyond this question, which is admittedly an essential one, the point that interests us here involves the observation that, following the Bolzano-Husserl definition of logically analytical truths, a proposition of the type ‘A surface cannot be at once red and green’ cannot be considered analytical either in the strict or in the broader sense. Indeed, if we replaced ‘green’ by ‘round’ or ‘scarlet’, this proposition becomes false. And seeing that, according to certain commentators including Simons himself (Reference Simons and Simons1992: 374 s.), it would not be plausible to ‘claim that the conceptred includes the characteristic not green’, assertions of this type could not even implicitly beanalytical, contrary to assertions by essential predication (‘A bachelor is an unmarried man’). It is therein that resides the strength of the Husserlian thesis by which there exist a priori (or materially a priori) synthetic propositions, and as a consequence, the weakness of the critique addressed towards him by Schlick.

B. Is there a material a priori? If we consider those a priori which are said to be ‘material a priori’ in the light of the whole of the four elements which we have just pointed to, it seems to me that we come up against a particularly thorny epistemological problem. We find ourselves in effect faced with an alternative where each option presents elements of difficulty.

(i) The first option involves following Simons where he explains that material a priori cannot be reduced to analytical truths of the Bolzano-Husserl type, but declaring in the end for the plausibility of the affirmation according to which ‘the concept red includes the characteristic ‘not green’ (whereas Simons denies any plausibility to that statement). Thus, we treat the statement ‘Nothing can be simultaneously red and green’ as a judgement (even an assertion) whose negation must be considered as an expression of conceptual (or linguistic) discordance, a case of displaced usage of the concepts (or of the language). But in that case, this assertion, without being analytical in the Bolzano-Husserl sense, shows up as being implicitly analytical: its truth, on the model of that of the statement ‘Bachelors are unmarried men’, depends on the concepts, or meanings, associated with the terms concerned. The nature of these explains the a priori character of the statement as well as its universal and necessary validity with respect to objects of experience.

Such a position could be based on those interpretations of Husserl which attribute to phenomenological analyses the merit of showing that the constraints arising from the grammar and semantic competence of speakers are rooted in the fact that ‘any concrete object which offers itself to our experience’ shows essential, structural properties (Reference Costa, Franzini and SpinicciCosta, Franzini, & Spinicci, 2002: 75). Now, posing things in this manner means affirming two things, of which one militates in favour of Husserl, the other of Schlick.

Following the first affirmation, we cannot understand the meanings which we attribute to certain expressions (for example, the term ‘red’) without referring to the eidetic intuition on which repose the concepts which these expressions reflect (for example, the concept <red>).

The principal merit of phenomenological analysis would thus be less that of having re-established the synthetic a priori refused by the neo-empiricists than of having made clear the limitations of another thesis of neo-empiricism: being the idea that the linguistic orientation of philosophy should allow for the rapid resolution of the problems of the formation of concepts and the validity of judgements and assertions by appealing to ideas – such as those of definition, of convention, or of semantic-linguistic rules – which are considered as not needing any particular philosophical clarification.^{Footnote 8} Having accorded that to Husserl, we must equally affirm – this time in accordance with Schlick – that if the meaning of the word ‘red’ derives its source from the content of the concept <red> where that concept is, for want of a better term, ‘structural’ and not simply classificatory, then the negation of assertions held to be materially a priori (such as ‘A thing cannot be at once red and green’) expresses less a discordance at the level of beliefs than an inadequate comprehension of the concepts involved and a slanted use of the words concerned. And this even if one does not lend to these truths a logico-analytical nature, whether in the strict or in the broader sense.

(ii) The second option, on the other hand, involves maintaining that concepts (and predicates) such as <red> are purely classificatory, in the sense that they serve simply to group together things which are red in contrast to things which are green without any structural implication as to the relationships between the fact of being red and of being green. In this case, the judgement (or assertion) by which a thing cannot be at once red and green is certainly one of synthetic nature: but declaring its syntheticity signifies admitting that this judgement (or assertion) may be denied without contradicting the content of the concepts which it contains (in other words, without violating one or other semantic rule). It is thenceforth logically possible, that is, non-contradictory, even though it may be factually false, that there does exist something which is at once red and green, with ‘red’ no longer indicating a quid which would imply reference to a conceptual structure embodied within the concept <red>. A discordance on the truth value of this judgement (or assertion) would therefore no longer be a conceptual (or linguistic) discordance, but a discordance of belief.

From this, however, there arises a truly taxing problem for the upholders of the material a priori. Indeed, if – despite what we have just said – those who maintain that the truth of these assertions is guaranteed a priori and that, as a consequence, these assertions possess a universal and necessary validity as concerns the objects of experience, they find themselves confronted – as Schlick did not fail to recall – by an obstacle identical to that which Kant had encountered when he had to explain the possibility of synthetic a priori judgements by means of a set of transcendental arguments (Reference Schlick, Schlick and WaismannSchlick, 1969a: 21–23; Reference Schlick, Mulder and Van de Velde-Schlick1979, II: 162–164). In other words, even the phenomenologists would be obliged to reflect upon the manner in which a synthetic or material a priori might be explained and would have to work hard to justify its validity by developing appropriate forms of argument.

The problem of such a justification became even more acute in the light of scientific developments of the nineteenth and twentieth centuries. We know that these led to the crisis of the Kantian synthetic a priori and cast renewed doubt on the possibility of judgements or assertions which, while having a given content, may enjoy a universal and necessary validity for objects of experience, putting them beyond contradiction by any potential development of our knowledge. To my way of understanding, this obstacle cannot be eliminated or by-passed by insisting on the gaps in the empiricist theory of concept formation or on the difficulty of placing phenomenological propositions among the logico-analytical assertions in any sense of analyticity which is acceptable or at least shared. Despite the limitations of the critique of Schlick and the other neo-positivists, this critique was grounded in serious philosophical reasoning, the same as that which had led these philosophers to deny the Kantian synthetic a priori. The significant epistemological point consists therefore in knowing whether Husserl's referral back to the intuition of essences, to eidetic variation in the imagination, and to passive synthesis permits the assurance to these claimed materially a priori judgements (or assertions) a validity which is a de iure and not simply a de facto vality, which concerns the objects themselves and not simply our phenomenological experience of these.

This had been doubted by Schlick. Subsequent to him, this doubt has been maintained by many others, myself included (Reference ParriniParrini, 2002, Reference Parrini and Lanfredini2006).^{Footnote 9} If, furthermore, account is taken of the critiques addressed by Quine around the distinctions between a priori and a posteriori and between analytical and synthetic, one is obliged to observe that consequently the situation becomes more complex, whether relating to the Schlickian or Husserlian solution. I believe that, to re-establish these distinctions, it is indispensable to link them back to the logical structure of epistemic justification, so as to end up with a relativized or contextualized a priori which is able to vary along with the variation of justification contexts. But that means removing from the a priori any absolute character. It comes down to integrating the a priori within a network model which attributes a contingent character to any asseverative element and seeks to maintain the distinctions which we have just mentioned (as well as that between statements which are protocols and those which are not, as we shall see) in consideration of the different functional roles exercised by assertions in the different contexts of epistemic justification and in the different scientific theories (Reference ParriniParrini, 1998). Thus, even well-tested factual assertions (as the history of science teaches us!) may be raised to the rank of contextually a priori definitions or principles. On needs think only, for example, of the principle of the constancy of the speed of light in the special theory of relativity (Reference ParriniParrini, 2011: 112–117, 127–128).

Objectivity

The nature of the a priori, and notably the status of the synthetic a priori, marks one of the fundamental differences between phenomenology and the approach of Kant. Out of these emerge two ways of conceiving knowledge and the constitution of objectivity which, despite having certain aspects in common, remain sharply distinct. Let us take the points of convergence first: first of all, both Kant and Husserl hold in common the idea that the cognitive process requires the co-operation of the intuition and the intellect (or understanding). This thesis becomes one of the principal elements of discord between Husserl and Schlick (whose critique of Husserl is, besides, not free from some incongruence). Whereas for Schlick, intuition has not to do with knowledge but the ‘enjoyment of life’, for Husserl as for Kant two elements come together in knowledge, one of which is conceptual, the other experiential and intuitive. It is true that, in Schlick as in Husserl, knowing something requires an operation on the level of recognition; but contrary to Schlick, Husserl thinks that recognition implies a reference not only to the concept, but also to an intuitive filling. As Roberta Lanfredini wrote (Reference Lanfredini, Parrini, Salmon and Salmon2003: 50–51), although ‘identification with a sensation of red is not a sufficient condition for the determination of an intuitive act, it is in any case a necessary condition’ for it. One cannot identify a red object without experiencing ‘what it is like’ to see something red (to use an expression of Thomas Nagel). That is why ‘it is possible to distinguish in phenomenological terms the determination or discrimination of a perceptive object (the result of a simple act of perception) from the identification of an object (the result of the conjunction of a meaning act and a perceptive act)’ (Reference Lanfredini, Parrini, Salmon and Salmon2003: 50–51, 49 – cf. Reference ParriniParrini, 2002: ch. 2, Reference Parrini, Cimino and Costa2012: 96–104, Reference Parrini2014).

Kant and Husserl both use the concepts ‘constitution’ and ‘transcendental’. For them, the fact of using these notions shows that the possibility of evoking an object in itself which is divorced from any link with the phenomenal experience of the cognizing subjectivity is excluded: and this, whether this link is established by a reference to the diversity of the sensible manifold (as in Kant) or it is anchored in the ‘motivational linkages’ which tie our judgements to experience (as in Husserl; cf. Reference Lanfredini and LanfrediniLanfredini, 2004: 1712, Reference Lanfredini and Lanfredini2006a: 91–93). Obviously, the experience to which reference is made in this latter case is not that of the Elementarerlebnisse of the Carnapian Aufbau: it is rather an experience defined according to the terms of the theory of intentionality, thus through a structural or analytical distinction between hyletic contents and intentional morphè, between noetic elements and noematic elements.^{Footnote 10}

Lastly, one finds in Kant as well as in Husserl the idea of ‘object in general’. Kant introduces it in a famous passage of the Critique of Pure Reason: ‘it is easy to see that that object can only be conceived as something in general = x’, adding that the unity constituting the concept of an object ‘cannot be anything but the formal unity of our consciousness in the synthesis of the manifold in our representations’ (Reference Kant, Guyer and WoodKant, 1968: A104–1905). As for Husserl, several commentators on phenomenology, without necessarily being in agreement over the manner of understanding the notion of ‘noema’, nevertheless detect in Husserl's description of the (cognitive) noema a connection back to the idea of an object in general x.^{Footnote 11} That is why some of these commentators declare that also in Husserl ‘existence’ is a ‘thetic predicate’. For both philosophers, existence does not refer to a concept in the form of a real predicate (being a concept which can be added to the concept of a thing), but in the form of a that is ‘merely the positing of a thing’ (Reference MohantyMohanty, 1996: 29).

A non-negligible difference nevertheless distinguishes Husserl from Kant. It is often highlighted that the theory of passive synthesis and the material a priori inasmuch as an a priori embedded in phenomenological experience leads Husserl to maintain that objective validity and truth require that ‘the implications of the noema are not contradicted by the unfolding of the course of our experiences’ (Reference MohantyMohanty 1996: 29). On the other hand, Kant's transcendental constitution is marked not only by the integration of experiences into a unitary objective experience, but also by a certain constructivism; that is, this integration supposes certain rules concerning the possible experience.^{Footnote 12}

In Kant, cognition requires, as well as an empirical material intuitive component, a conceptual structuration linked to forms of thought which emanate from the spontaneity of the intellect. It is thus that we can reach the level of cognitive objectivity and of objects of experience (or phenomenal objects), in accord with the supreme principle of a priori synthetic judgements (‘the conditions of the possibility of experience in general are at the same time conditions of the possibility of the objects of experience themselves’ Reference Kant, Guyer and WoodKant, 1968: A158 = B197). The fact of having established this principle by appropriate argumentations (the transcendental expositions of space and time, the deduction of categories, and the demonstrations of the principles of the pure understanding) leads Kant to believe that he has established how a priori synthetic judgements are possible. He considers that these judgements are possible because they arise from the forms of the sensibility and the intellect which initiate the unification or synthesis of the sensory material: it is this unification which constitutes our knowing of objects of experience, that is, of phenomenal objects as known objects and distinct from the things in themselves or considered by themselves. I am speaking of constitution of the knowing of objects of experience to guard against certain readings of the Kantian text which, going beyond legitimate interpretive divergences, falsely interpret transcendental gnoseology in the idealist-subjectivist sense: as if Kant were maintaining the frankly ridiculous thesis that it is our reason or our mind which produces in all and for all the object of knowledge. Which proposition, let it be said in passing, generates forms of realism which are philosophically quite rudimentary.^{Footnote 13}

At all events, beyond whatever differences in reading there might be, the constructive character of the Kantian constitution of cognitive objectivity remains firmly established. One need only think of the way in which Kant traces the boundary between real perception and illusory perception, between hallucination and reality, between dreaming and wakefulness. All these distinctions are established according to a conception of cognition as a synthesis of formal elements (of subjective origin) and material elements of empirical derivation: but one must emphasize the fact that the first proceed from a spontaneous activity of the intellect, that is to say, of the cognizing subjectivity. As one reads for example in ‘Observation III’ of the Prolegomena, the distinction between dreaming and truth isn’t ascertained by the nature of the representations in question, for they are the same whether or not one is dreaming, but by their inter-connections according to the rules [Regeln] that bring representations together under the concept of an object and settle whether or not they (the representations) can coexist in a single experience (Reference KantKant, 1995: 290).^{Footnote 14}

I shall come back to the fact that this refers in Kant to an alethic evaluation of judgements which cannot be based solely on intuition. Let us restrict ourselves for the moment to recalling that scientific developments post-Kant have certainly led the neo-empiricists to contest the possibility of norms and principles of judgement given once and for all: but these latter have nevertheless not regressed into a naïve empiricist conception which would neglect the role of concepts and principles which allow the organization of the material of experience to constitute a cognitive objectivity. One need only think of what Schlick and Reichenbach maintained, occasionally in disagreement, occasionally in agreement, on the notion of co-ordination, on the principles of constitution and co-ordination as well as on the co-ordinative definitions or conventions; as many positions which, despite their differences, integrate constructivist-type motifs of Kantian and conventionalist origin. One need also only think of the doctrine of the physical world of the Aufbau: besides his refusal of the synthetic a priori, in this doctrine Carnap invokes methodological principles allowing the assignment of sensible qualities to spatio-temporal points-instants, such as to constitute or construct, as Quine states in ‘Two Dogmas of Empiricism’, ‘the laziest world compatible with our experience’.^{Footnote 15}

To conclude on this topic, I would like to emphasize that the subjective and constructive element of cognition has retained its full importance, even after the crisis of the neo-empiricist doctrine. Epistemological reflection on themes like the empirical sub-determination of scientific theories and the holistic character of the relationship between theory and experience has in effect led to the thesis that this relationship is of a triadic nature. In other words, it is made up of (1) judgements submitted to empirical examination, (2) contents of experience such as have been codified in certain protocol statements, and (3) a series of judgement presuppositions of a character that is linguistic (analytical a priori), theoretical (relativized or contextual synthetic a priori), and methodological (simplicity, compactness, familiarity, etc.) (Reference ParriniParrini, 1998: ch. 2).

Judgement

All this leads us to the question of judgement. We have seen that, for Kant as well as for Husserl, knowing implies integrating both the intuitive and conceptual aspects. For Kant, this integration must avoid the situation where the truth or falsity of a judgement (or, at all events, the acceptance or rejection of this latter) violates norms, or principles, of the faculty of making judgement. A sort of equilibrium – which I would call ‘reflexive equilibrium’ – must be achieved between certain abstract forms of thought, judged as valid a priori, and the contents of experience. Thus, in the Kantian tradition, the ideas of cognitive objectivity and of truth acquire an axiological dimension which precisely links back to the respecting of the norms and principles of judgement. However, a very interesting debate has arisen on this subject, whose protagonists have been, on the one hand, the Neo-Kantian Heinrich Rickert, who has criticized the phenomenological theory of intentionality, and on the other, Martin Heidegger: the latter, in the course of lectures on the concept of time in 1925, dismissed Rickert's critique, aligning himself with Husserl and laying the bases both for the theory of truth which he would subsequently develop in Sein und Zeit as well as the evaluation of Kant's alethic conception which is contained in Kant and the Problem of Metaphysics.

In order to ensure clarity, I will begin at the end of the argument and proceed in reverse order. Let us take the famous example in Sein und Zeit in which Heidegger illustrates his idea of truth as aleitheia or uncoveredness: a man determines to check whether it is true or false that the painting on the wall behind him is hanging crookedly. The man, says Heidegger, will turn towards the wall to establish the truth or falsity of the judgement:

This way of understanding truth links back to an incomplete and in part erroneous interpretation of the Kantian conception of truth. Contrary to what the Neo-Kantians have affirmed, Heidegger does not see in Kant any going beyond the traditional (and to the Neo-Kantians, naïve) doctrine of truth as being a copy or adaequatio intellectūs ac rei. To the contrary, Kant would retain that doctrine (Reference Heidegger and StambaughHeidegger, 1976: 285–286 = 215; Reference Heidegger1991: 12–13). However, Heidegger neglects the fact that Kant accepts this classical doctrine of correspondence exclusively as a nominal definition of truth, in that it says nothing about the criterion for deciding about the truth or falsity of an apophatic judgement. As for the criterion of truth on the other hand, Kant openly declares that it must reflect both the material conditions and the formal conditions of knowledge – with these latter being expressed in their turn in logico-analytical conditions (the principles of general logic and of the analytical judgements assuring the non-contradictory character of the judgement) and in logico-transcendental conditions (the set of synthetic a priori principles which allow the judgement to lay claim to an objective validity) (Reference Kant, Guyer and WoodKant, 1968: A57–60 = B82–85; cf. Reference ParriniParrini, 2011: ch. 1, ch. 2 § 2).

What I wish to emphasize in this regard is that both this Heideggerian conception of truth and the interpretation of Kant that is linked with it are rooted in the way in which Heidegger in his 1925 lectures opposed Rickert. In the course of those lectures he shows that he is plainly aware of the fact that the Husserlian theory of intentionality carries a distinction between the sensible elements of perception and their objectual reference (the sensation of red and perceived red, sounds heard and the perception of the soprano's singing). This induced him on the one hand to reject (correctly) the critique of Husserl by Rickert, but on the other hand to reject (wrongly) the whole Kantian and Neo-Kantian theory of objectivity. To be more precise, Heidegger opposed Rickert's idea according to which ‘the knowledge object is a value’. Trying to demonstrate the absurdity of the critique that had been directed at Husserl, Heidegger was already using an example very close to that of the aforementioned painting of Sein und Zeit:

That is no doubt true and we will not revisit the question of whether and to what extent Rickert may have incorrectly interpreted Husserl's conception centred on intentionality, thus on the refusal of any comparison between objects themselves and mental representation of the objects. There is no doubt that the truth of the judgement about the painting or the chair is grounded in the conformity of this judgement with the painting or with the chair; there is equally no doubt that that towards which one's intention is directed in the perception is the painting or the chair. Nevertheless, as we have seen with respect to the distinction between real and unreal perception, between dream and reality, the establishment of this adequation or conformity does not derive solely from what is given in perception (one might think of the famous segments of the Müller-Lyer illusion which seem to be of different lengths but which in fact are identical in length). And if that holds true for judgements on the way paintings are hung, on the number of legs chairs have, and on the length of line segments, what might be said about judgements on viruses or neutrinos! One must certainly recognize that for the purposes of evaluating truth or falsity, or simply for the empirical adequation of any synthetic judgement, it is necessary to specify the links, however loose they might be, which tie, in the contexts of justification, judgements submitted to the test of experience to ‘protocolary’ empirical judgements; but we must perforce recognize just as distinctly that certain empirical judgements, each time in a determined control context, will have to be accepted or rejected from an ‘evidence’ of perception alone, from what is provided immediately in the perception, in the manner of the interpretation Heidegger offers us in relation to the painting or the chair. But all this does not prevent the validity of judgement, even a judgement such as that concerning the chair or the painting, from not arising solely from what the perception provides, as this could turn out, for the most diverse of reasons, to be deceptive or illusory.

In summary, perception cannot be invoked as the sole and unique criterion for establishing which of my perceptions is real and which is deceptive, for I find myself in both cases faced with something which is provided in the form of perception (hence the sceptical arguments!). The truth criterion for judgements cannot be reduced exclusively to that which is given by perception. To the contrary, this criterion must incorporate an evaluation of the congruence of the judgement with a set of relatively a priori principles which form the structure of a set of judgements which is the most comprehensible, coherent, simple, and compact possible. This is indeed what Kant had affirmed after his own manner when he had stated that

The validity of the judgement about the painting which seems to be hanging crookedly on the wall or on the chair which had or did not have four legs does not depend solely on what is provided to the sensory intuition (to the experience); it depends also on the set of possible judgements to which this judgement may be related in holistic fashion (and which touch, for example, upon instruments or even operations of measurement). And what is valid for common sense judgements is a fortiori equally valid for science, that is, for the complex, abstract, and very sophisticated interconnected array of judgements in which our scientific knowledge is condensed and summarized. There is no doubt that this rich and multi-linked structure requires in one manner or another to be grounded in judgements of experience which, according to the contexts of epistemic justification, are accepted or rejected (if only temporarily) on the basis of intuitive experience. There is also no doubt, at least from my point of view, that these intuitive experiences must be understood and interpreted in accordance with the Husserlian theory of intentionality. But that should not occlude an awareness that the evaluation of other judgements depends also on hypotheses of a linguistic, theoretical, and methodological nature which constitute the axes along which epistemic justification is laid out (epistemic relativism and a relativized a priori; see Reference ParriniParrini, 1998: ch. II).

None of the elements which structure our knowledge – according to a model which in my opinion should be of a reticular nature (Reference ParriniParrini, 1998: ch. III) – can lay claim to an absolute validity. Even the protocol judgements (or assertions) may be in one particular context considered as baseline judgements to be accepted or rejected according to experience (thus, used for the empirical control of other judgements or assertions), but put back into question in another context in favour of other hypotheses and judgements (or assertions) which are in their turn considered to be of a baseline nature in this different context. A phenomenological theory of the knowledge object ought to take into account, it seems to me, this aspect which is characteristic of ordinary knowledge and, especially, of scientific knowledge. In this latter case, a phenomenological theory should even seek to explain what we have learned about the historical dynamic of science, that is its aspects of continuity and progressive broadening along with its apparently discontinuous aspects of change and rupture.

Whence, in my view, the importance of a synthesis or integration between phenomenology, the Kantian tradition, neo-empiricism and post-positivism to bring to fruition a general theory of science and knowledge. Undoubtedly we must show prudence if we are to advance on this terrain. The phenomenological approach to epistemological problems on the one hand and the neo-empiricist approach as heir of the Kantian (and conventionalist) tradition on the other have been animated by profoundly divergent interests, which it seems led the phenomenologists to elaborate a theory of knowledge as intentional act and the neo-empiricists to set out the first pointers to the contemporary philosophy of science through the interpretation of the relationship between theory and experience in terms of cognitive co-ordination. But what we have learned and understood over the last few decades shows that these two approaches were developing, each within their own sphere, themes and perspectives whose results, rather than simply just allowing, were effectively demanding to be reciprocally integrated together as parts of a much broader conception. That is why we will come particularly to appreciate the efforts of certain authors such as the Austro-American Felix Kaufmann and the Italian Giulio Preti, who from the end of the 1940s perceived the possibility of this integration and who strove through their work to bring this about.

Translated from the French by Colin Anderson