Academia.eduAcademia.edu
MEINONG ON MEMORY∗ Fabrice Teroni Department of Philosophy Geneva University Meinong’s early essay, Zur erkenntnistheoretischen Würdigung des Gedächtnisses1, provides, despite its brevity, a very important discussion of mnesic phenomena. In this paper, I investigate some points I believe to be of particular interest: some of them are briefly broached by Meinong, whereas others form an important part of his argument. Moreover, I will connect the discussion with more recent concerns in analytical philosophy. This presentation is structured in the following way. In section (1), I present Meinong’s definition of memory, and contrast it with more recent conceptions. In section (2), I present his epistemological discussion of memory judgements, which is further investigated in section (3), where I discuss the notion of presumptive evidence via the criticisms it elicited from Franz Brentano. At the end of this presentation, we will understand why Meinong’s essay is a landmark in the philosophical investigation of memory. (1) WHAT IS MEMORY? Meinong’s interest in memory is primarily epistemological, his goal being to take possession of this area in the name of epistemology2. According to him, the problems surrounding the epistemological assessment of memory have been neglected, philosophers being more interested in broad questions than in elucidating more mundane, though fundamental, data of knowledge3. But of course, even though Meinong’s goal is of this nature, he must start with a conception of memory providing the foundations for the discussion of its epistemological status. What form does it take? In this text, Meinong’s remarks are quite sketchy, but he nevertheless states some fundamental theses we can contrast with many other ∗ Thanks to Christoph Hoerl for many helpful comments. MEINONG, Alexius, Zur erkenntnistheoretischen Würdigung des Gedächtnisses, in HALLER, Rudolf, and KINDINGER, Rudolf (eds.), Gesamtausgabe, Band II, Akademische Druck, Graz, 1969, translated as Toward an Epistemological Assessment of Memory, in CHISHOLM, Roderick, and SWARTZ, R.J. (eds.), Empirical Knowledge, Prentice Hall, 1966. References are given first to the translation, and to the original in parentheses. 2 MEINONG, Alexius, op. cit., p. 254 (p. 188) 3 MEINONG, Alexius, op. cit., p. 254 (p. 188) 1 2 MEINONG ON MEMORY accounts. His first point consists in distinguishing memory from presentations. A presentation occurs for instance when a painter imagines a scene before drawing it: he has a presentation of something, but this ‘something’ exists only in [his] thought[…], it is nothing but the content thereof.4 Memory differs from presentations because it is, to put it briefly, assessable as true or false, as justified or unjustified, contrary to presentations. They nevertheless play a crucial role: even though they are not by themselves instances of knowledge, they can provide its basis. Accordingly, Meinong understands memory as judgements based on presentations: What is added is the clear and, here, very essential conviction [Überzeugung] of the person remembering, that the image in his memory relates to an actual experience.5 This conviction is grasped by Meinong as the occurrence of a judgement which is, of course, assessable as true or false, and as justified or not6. Memory is a form of knowledge, a judicative act linked with a specific presentation. Do these presentations have a specific nature and differ, for instance, from imaginative presentations? Meinong does not provide a clear answer here, and one can only gather the core of his position through some negative remarks. For before stressing the fundamental role of judgement, he notes that the distinction between memory and presentation cannot be reduced to the simple fact that the contents originated in the earlier experiences of the person remembering, or the knowledge of such origins gained from psychological investigation –for both might also be present in the artist who creates out of ‘pure fantasy’.7 This may be the case, but it is important to note that Meinong does not, and should not, say that these properties play no role in an account of memory. If it is a kind of judgement, and is moreover based on presentations, then presentations themselves must satisfy certain constraints. For if memory judgements were made on the basis of presentations that do not, in a certain way, originate from earlier experiences, then the epistemological status of memory would be problematic, and the difference between memory and (at least some) imaginative presentations would disappear. Meinong should not 4 MEINONG, Alexius, op. cit., p. 254 (p. 189) MEINONG, Alexius, op. cit., p. 255 (p. 189) 6 MEINONG, Alexius, op. cit., p. 255 (p. 189 in fine). Thus, Meinong does not make reference to something more primitive than judgement, contrary to Russell. See RUSSELL, Bertrand, The Analysis of Mind, Routledge, 199. This very important point is only superficially broached in what follows. 7 MEINONG, Alexius, op. cit., p. 255 (p. 189) 5 FABRICE TERONI 3 reduce the difference between memory and imagination to the occurence of a judgement directed to the past, but his epistemological interests lead him to stress its fundamental role. His position is compatible with many specific conceptions of their nature, and should not be assimilated to those of philosophers who have stressed the dispensability8 and generality9 of memory ‘images’ which are particularized with the help of judgements. For Meinong simply says that to remember is in part to judge, not that memory presentations can be of the same sort as what occurs in any case of sensory imagination. Even if it were true that identical presentations can also be part of imaginative acts, this should not obscure this fact10. Moreover, as I briefly point out below, in order to reach his specific epistemological position, Meinong should be more specific on the nature of memory presentations. But what is the source of Meinong’s stress on judgements? It is to be found in his classical methodology: he studies memory from the first person point of view. And, as he argues, it would certainly never occur to anybody to say that he remembers this or that, if he lacks such a conviction.11 This is no doubt the case, but one could easily question the hidden argument: why should the fact that one is ready to self-attribute mental states have any role to play in their definition? Even if no one would say that he remembers were he not convinced that something occurred in the past, why should this play a central role in our conception of memory? For note that there exist many situations, both intra- and interpersonal, where we attribute memory at a time when no judgement occurred. A case of the first type is the following. At t1, John has a memory presentation, but does not make a judgement in the past tense about what is presented because he has reasons to believe that such cannot have been the case. At t2, he comes to know that his former reasons were completely unfounded and says ‘I remembered this, but was not at the time inclined to believe that such was the case’12. Meinong should have used inclinations to believe, and not judgements, to escape this problem. Case of the second type are more problematic and occur whenever someone shows us, in a certain way, that the past has a very specific influence on his present doings, 8 On the first point, see for instance ZEMACH, Eddy, A Definition of Memory, Mind, 77, 1968 ; for dissenting views, see HOERL, Christoph, The Phenomenology of episodic Recall, and MARTIN, M.G.F., Episodic Memory as retained Acquaintance, both in HOERL, Christoph, and McCORMACK, Theresa (eds.), Time and Memory, OUP, 2001 9 See EVANS, Gareth, The Varieties of Reference, OUP, 1982, p. 268 10 My caution is explained because, on some possible accounts of memory presentations, they are different from imaginative presentations independently of judgement. Such is the case if, for instance, objects are primitively presented as having occured previously. 11 MEINONG, Alexius, op. cit., p. 255 (p. 189) 12 See section (2). 4 MEINONG ON MEMORY without his realizing that this is so. We say for instance that he remembers a specific event without believing that it occurred13. So, Meinong’s stress on judgements depends on a questionable methodology, and has some problematic consequences, but is basically sound because of his central concern: even if it can be argued that to remember is not to judge, memory judgements nevertheless take central stage when epistemological problems are investigated. For only judgements and beliefs can be evaluated as justified or not. An important thesis is Meinong’s restriction of memory to judgements occurring in specific contexts, namely when specific presentations occur. He consistently uses the term ‘Erinnerungsbild’, with some occurrences of ‘Erinnerungsdatum’ (which are sometimes, correctly in my opinion, translated also as ‘memory images’14). It seems therefore that, according to Meinong, memory is always quasi-sensory, or experiential, for he would probably have used the more neutral term ‘Vorstellung’ instead of ‘Bild’ if he wanted to cancel this implication. This is important, and in sharp contrast with many contemporary conceptions of memory. For it is now a psychological and philosophical commonplace to cite different forms of memory, the usual list comprising at least procedural, propositional or semantic, and personal or episodic memory. Procedural memory occurs for instance when Mary remembers how to swim, something philosophers use to refer to as procedural knowledge or know-how. Examples of propositional or semantic memory are: John remembers that Napoleon crossed the Alps or Michael remembers that 2+2=4, whereas episodic memory is restricted to events the subject has witnessed. This conception of the scope of memory is liberal, and Meinong is opposed to this idea: memory is restricted to the episodic species, which he moreover understands as experiential, in the sense that to remember is to enjoy phenomenologically rich states of mind. The best way to flesh out the traditional talk of ‘memory images’ consists in saying that to remember can be, for the subject, as if he was hearing or seeing. This leads to a restricted, by contrast to a liberal, position on the scope of memory: one remembers only when one judges on the basis of such specific states of mind. Propositional memory is fundamentally different, being nothing else that the retention of a judgement. But Meinong’s position remains unclear, for even though he constantly uses the term ‘memory image’, at one point he notes that we remember thoughts and feelings, judgements and desires.15 In some of these cases, it seems implausible to extend the conception of memory just mentioned, for a judgement can be retained without 13 For a classical example, see MARTIN, C.B. and DEUTSCHER, Max, Remembering, The Philosophical Review, 75, 1966 14 See for instance p. 262 (p. 199) 15 MEINONG, Alexius, op. cit., p. 256 (p. 191) FABRICE TERONI 5 any memory image being required. A plausible interpretation consists in distinguishing judgements that I desired or that I judged that p in the past from memory judgements, which are constitutively accompanied by presentations of the contexts in which these acts have taken place. In this sense, to remember a judgement is to remember judging that p, thereby grounding a contrast between preserved judgements and memory judgements. Another option starts by arguing for the sensory character of these acts, to encompass them in this way in one’s talk of memory images, but this strikes me as far less plausible than the first position. For reasons that I can only briefly mention here, I think Meinong is perfectly right in his restricted conception of the scope of memory, for many fundamental properties distinguish the phenomena he understands under this label from cases of propositional or semantic memory, which are therefore best grasped as completely distinct. Among them, one can cite the phenomenological dimension just mentioned, the specific modes of thought characteristic of memory (memory stricto sensu seems to allow for demonstrative thinking, contrary to propositional memory16), and a specific form of intentionality. More importantly for Meinong’s own purpose, the epistemological assessment of memory, it is very plausible to argue that memory and preserved judgements have completely different epistemic structures. There are therefore many reasons to accept Meinong’s restricted conception of memory judgements. Another important position held by Meinong follows from this restricted conception of memory and his specific analysis of mnesic phenomena, namely the fact that he distinguishes between different types of judgements made on the basis of memory presentations. Let us go back to the citation at the beginning of page 2, where Meinong notes that to remember is to be convinced that the image relates to an actual experience. He immediately clarifies what he means: i.e., that [the subject] is thinking of something that actually happened, and not of something that (…) he arbitrarily or accidentally thought up or invented.17 The relation between memory and experiences is best addressed by two questions. First, and very importantly for those who stress the experiential character of memory, what is the relation of memory judgements to present memory experiences? Moreover, and contrary to the case of perception, there is a second question: what is the relation between memory judgements and past experiences? Let us investigate Meinong’s answers in turn. As regards the first, he notes that 16 17 CAMPBELL, John, Reference and Consciousness, OUP, 2002 MEINONG, Alexius, op. cit., p. 255 (p. 189) 6 MEINONG ON MEMORY while I am remembering, I can easily make a judgement about that which I am remembering, without, at the same time, making a special judgement about the remembering itself.18 Typical memory judgements occur, according to him, when one judges that there was a wonderful party at John’s place last summer, or that this lake was wonderful. And cases (…) where the remembering itself forms the content of the judgement (…) must be regarded as more complicated and, in this respect, as secondary constructs, for here the simple act of memory is supplemented by reflection on itself.19 We must distinguish between memory acts that are, as we saw above, judgements made on the basis of specific presentations, from more complex judgements about this act.20 As regards the first question, Meinong’s answer is therefore that for a subject to remember, i.e. to judge about the past, he is not required to judge that he enjoys a specific memory experience: memory does not depend on these further achievements. Meinong’s position can be interpreted in two ways: either as saying that to require this achievement is to mistake dispositions and occurrences (S is simply disposed to judge that he has a memory experience when he makes a memory judgement), or that the disposition itself is not even required in order to form memory judgements. What about the second question, the relation of memory to past experiences? The issue is more tricky, for Meinong’s remarks are issued in order to bypass the problem of realism. He argues for a reconciliation of what I take to be idealism and realism by stressing that we should give a neutral account of memory, leaving the problem of knowing whether it answers to any independent reality aside. Here is his argument: (1) (2) (3) One can only remember what one has experienced. One can only experience what goes on within oneself. Therefore, one can only remember what went on within oneself. The conclusion is more precisely that 18 MEINONG, Alexius, op. cit., p. 255 (p. 190). He similarly refuses to conceive memory judgements as sophisticated judgements of correspondence, see op. cit., p. 259 (p. 195). 19 MEINONG, Alexius, op. cit., p. 255 (p. 190) 20 This is in sharp contrast with Brentano, according to which each state of mind has itself as secondary object, something he describes as eigentümliche Verfleckung. See BRENTANO, Franz, Psychology from an Empirical Standpoint, Routledge, 1995, and JACQUETTE, Dale, Brentano’s Concept of Intentionality, in JACQUETTE, Dale (ed.), The Cambridge Companion to Brentano, CUP, 2004, p. 101. FABRICE TERONI 7 we can actually and directly [unmittelbar] remember only the data of the mental life (…).21 Anything else is remembered « on the basis of such data », as Meinong puts it. His strategy seems to consist in discounting the problem of realism because an account of memory must in any case be given independently of its resolution: what follows clearly shows that he is happy to restrict the discussion to judgements about the subject’s former states of mind. Before proceeding, it is interesting to underline that Meinong grasps knowledge of what is external to the mind as problematic, contrary to knowledge of what has occurred in the mind itself. Here, he is close to Russell who, in The Problems of Philosophy, admits acquaintance with past sense-data, but not with anything else than sense-data22. We tend today to conceive knowledge of the past and of the external world as being equally problematic, as Russell himself was to argue later23. Let me return with Meinong’s restriction of memory to judgements about past mental states. First, note that this is in sharp contrast with Meinong’s own examples of memory judgements mentioned above, which are about normal external events or objects, and seem independent both of judgements about present and past experiences. If we follow Meinong, we have to evaluate this as a simplification: memory judgements are primarily directed at past experiences, and only secondarily at past external events. A memory judgement about a past external event is always made on the basis of a judgement about a past state of mind, and the interesting conception of memory judgements as unsophisticated vanishes. But this sounds implausible: we do not all the time judge that we had experiences, but more commonly that events occurred. Meinong’s position is only motivated by the argument sketched above, so we should question its credentials if we want to avoid this implausible consequence. It is helpful here to sketch two types of direct/indirect realism with respect to memory. First, memory can be said to be about experiences in the sense that memory presentations are presentations of former experiences. This does not imply anything with respect to the object of the judgement, and is compatible with direct realism. Second, one can argue that memory judgements are primarily directed at past experiences, that past experiences constitute their immediate objects. This is a form of indirect realism, parallel to the idea that we immediately perceive sense-data. As far as I can see, pointing to the fact that one can only remember what one has experienced (premise (1)) grounds the first position: if memory presentations depend on previous experiences, then the restriction of memory to what subjects have previously experienced is respected. This uncontentious claim should not be confused with the contentious one that the object of memory is a past experience. 21 MEINONG, Alexius, op. cit., p. 256 (p. 191), my emphasis. RUSSELL, The Problems of Philosophy , OUP, 1970 23 See especially RUSSELL, Bertrand, The Analysis of Mind, Routledge 22 8 MEINONG ON MEMORY Moreover, premise (2) is at best dubious, for the mention of ‘experience’ is undeniably unclear: in a certain sense, I enjoy experiences which may be said to go on inside of me24, but this is not to say that their primary object is something of the sort. To say this is not to adopt a reconciliatory position, but to refuse direct realism, and Meinong seems happy with the alternative: idealism or indirect realism. This is to omit direct realism about memory, which stresses that memory judgements can be immediately about past events or objects. Meinong neglects here important distinctions as well as his former mention of memory judgements directly about external events. Even if it is true that memory judgements are made on the basis of mental states, this does not mean that they are immediately directed upon a present, or past, mental state. Nor, I would add, is the fact that memory presentations depend on former experiences a ground for saying that judgements made on their basis cannot be directly about past external events25. As far as I can see, Meinong’s position is easily reconciled with direct realism once his argument is dismissed: there is room for a direct realist Meinongian position according to which memory judgements are directly about past events. Note also that his peculiar conception of memory judgements is not grounded in classical foundationalist fashion, i.e. by arguing that uncertain judgements depend on certainty, for judgements about past mental states are uncertain and Meinong is not, as we shall see, committed to this position26. His answer to the second question is nevertheless that memory judgements are directly about past experiences, but is inconclusively buttressed. If we modify his position in the way suggested, we can see how a fundamentally Meinongian conception of memory differs from an important tradition which conceives memory judgements as complex and reflexive. Thus, according to Locke, the repository of the memory signifies no more than this, - that the mind has a power in many cases to revive perceptions which it has once had, with this additional perception annexed to them, that it has had them before.27 For Locke, as for William James28, memory is complex, for it depends on the mind’s capacity to think about its own past experiences, an 24 Disjunctivists are prone to question a specific reading of this remark, see HINTON, J.M., Visual Experiences, Mind, 1967 ; SNOWDON, Paul, Perception, Vision and Causation, reprinted in DANCY, Jonathan (ed.), Perceptual Knowledge, OUP, 1988 ; CHILD, William, Vision and Experience : The causal Theory and the disjunctive Conception, The Philosophical Quarterly, 42.168, 1992 and Vision and Causation : Reply to Hyman, The Philosophical Quarterly, 1994 25 These remarks must of course be grounded in a specific conception of memory presentations and judgements that I cannot provide here. 26 See section (2) below. 27 LOCKE, John, An Essay concerning human Understanding, Book II, Chapter 10, § 2. Emphasis in the original. 28 JAMES, William, The Principles of Psychology, vol. I, Dover Books, 1950, p. 648 FABRICE TERONI 9 important part of what is called today a ‘theory of the mind’. This position has many contemporary advocates who stress that memory, or at least an important form of memory (episodic or personal memory) requires the subject to be able to think about his own states of mind: they are metarepresentational, or second-order thoughts29. I said above that Meinong’s own stress on the unsophisticated nature of memory judgements can lead, once his argument for indirect realism is cancelled, to a plausible account according to which they are independent on both judgements about present and past experiences. This Meinongian account differs radically from the positions sketched here: according to it, memory occurs when modest judgements are made on the basis of specific presentations, more complex states of awareness being no more than other judgements made on the same basis. Memory is conceived as a fundamental endowment of the mind, which can be sophisticatedly exploited once the relevant capacities are acquired, but which functions independently of them. In a nutshell, memory is a distinct first-level mental phenomenon, and there are no different types of memory, but various judgements made on the basis of memory presentations. In this, it is on a par with perception, on the basis of which we can make reflexive judgements (such as ‘I am standing in front of a church’), but which does not depend on the capacity to make them: more mundane judgements are already perceptual. For reasons that I have no time to discuss here, but which are closely linked with the fundamental distinctions I mentioned above, I think this Meinongian position to be philosophically sound, and to ground an appealing conception of memory. (2) THE JUSTIFICATION OF MEMORY JUDGEMENTS Memory judgements have a specific nature, but Meinong’s main interest lies in their epistemological credentials. In this section, I discuss the way Meinong argues for the specific justification he endows them with, the next being devoted to the way we have to unpack his position. My aim here is not to reconstruct fully his argumentation, but to convey its motivation, plausibility, and similarity with other positions. Meinong points first that memory judgements are made with a distinct claim to credibility (…).30 29 Important contemporary accounts of this type are PERNER, Joseph, Experiential Awareness, in SCHNEIDER, W. et WEINERT, F.E. (eds.), Interactions among Aptitudes, Strategies and Knowledge in cognitive Performance, Springer, 1990 ; PERNER, Joseph et RUFFMAN, Ted, Episodic Memory and autonoetic Consciousness, Journal of Experimental Child Psychology, 59, 1995 ; DOKIC, Jérôme, Une Théorie réflexive du souvenir épisodique, Dialogue, 36, 3, 1997 ; Is Memory purely preservative ? in HOERL, Christoph et McCORMACK, Teresa (eds.), Time and Memory, OUP, 2001 ; OWENS, David, A Lockean Theory of Memory Experiences, Philosophy and Phenomenological Research, 56.2, 1996. 30 MEINONG, Alexius, op. cit., p. 255 (p. 189) 10 MEINONG ON MEMORY When I remember and judge that the house by the lake was made of brown wood planks, I feel entitled to this judgement, something which is for instance not present if I judge out of the blue. The goal of epistemology is according to Meinong to attempt to determine whether this trustworthiness [Vertrauenswürdigkeit] is based on something which these judgements have in common with others that epistemologists have already investigated, or whether it is based on something which, due to its special nature, requires special consideration.31 His method consists in investigating the possibility to classify memory judgements in classical categories. To complete this task, Meinong uses two fundamental distinctions within the category of judgements, which can be summarized in the following matrix: Is the judgement based upon proof? What is the type of judgement? Yes Existential Indirectly evident existential judgement. Ex : the judgement that there were giant lizards on Earth. Relational Indirectly evident relational judgement. Ex : mathematical judgement derived from a proof. No Directly evident existential judgement. Restricted to judgements of inner perception. Directly evident relational judgement. Ex : judgements of comparison and compatibility. Meinong next discusses, in turn, the plausibility of classifying memory judgements in each of these categories. He provides two reasons against assimilating them to directly evident relational judgements: first, a claim to correspondence may exist with respect to memory, ie the subject can claim that his memory image corresponds to what was the case, but this is as inessential to memory judgements (or only connected with them in the majority of cases) as the above-mentioned reflection on the remembering process.32 To classify memory judgements as relational is to intellectualize them too much. Second, and more crucially, judgements of correspondence depend on memory, since the reality to which the present image is compared lies in the past and, according to Meinong, relational judgements are only possible if both terms of the relation are given, and not if one of these terms must, as a precondition, belong to the past.33 31 32 MEINONG, Alexius, op. cit., p. 256 (p. 192) MEINONG, Alexius, op. cit., p. 259 (pp. 194-5) FABRICE TERONI 11 To compare Mary’s party with his memory image, the past party must be cognitively accessible to John, and this can only be provided by memory. Thus, to understand memory judgements as relational is to be blind to their specific form, and to be caught in a vicious circle, since the only remotely plausible relational judgements with which they can be equated depend on them34. Are they directly evident existential judgements? Meinong restricts this category, in a Cartesian fashion, to introspection, and shows how implausible it is to identify memory judgements with introspective judgements stricto sensu (they do not have the form ‘this occurs now in my mental life’35). Moreover, the only alternative is to say that they are made on the basis of introspection, but with an eye to what occurred in the past, something as ‘what I enjoy now is identical with something I enjoyed in the past’: as Meinong notes, this is to come back to the former implausible tentative to grasp them as relational. Memory judgements are neither relational nor directly evident existential judgements, thus they are not directly evident. This indirect proof by elimination does not satisfy Meinong, who adds important comments on the directly evident. What has the property of being directly evident has the characteristic of imposing an absolute firm conviction.36 If we come back to the matrix above, we see that directly evident judgements are for Meinong indubitable, for how could I doubt that 2+2=4, or that I am enjoying such and such a conscious experience? This does not lead him to say that to be directly evident is to be absolutely convincing: this is one of its characteristics, and to equate the two would be to confuse psychology with normativity. For unshakable conviction is but too commonly present when no evidence, but simply prejudice, has taken root in the mind. What is absent in this case, but present when what is directly evident elicits conviction is, as I understand Meinong, the right to be sure. In these last cases, unshakable conviction is justified, or grounded: the intensity of the judgement corresponds to available evidence37. I do not know how he conceives the relation between direct evidence and firm conviction (is the relation necessary? what kind of necessity does it have ?), but these remarks suffice for my present purposes. For to stress this 33 MEINONG, Alexius, op. cit., p. 259 (p. 195) I take Anscombe to make essentially the same point. See ANSCOMBE, G.E.M., Memory, ‘Experience’ and Causation, in Collected Papers, Vol. II, Metaphysics and the Philosophy of Mind, OUP, 1981, p. 126 35 This remark is also grounded in by distinction between sophisticated and modest judgements discussed above. 36 MEINONG, Alexius, op. cit., p. 260 (pp. 196-7) 37 This is in contrast with SCHUBERT-KALSI, Marie-Luise, Meinong’s Theory of Knowledge, Martinus Nijhoff, 1987, who opts for a more subjective reading of evidence. One thing is clear : Meinong distinguishes here evidence from conviction. 34 12 MEINONG ON MEMORY characteristic of directly evident judgements grounds another, and this time direct, reason why memory judgements cannot be of this type: it is indeed very common to doubt the deliverances of memory. Meinong has thus plausibly eliminated three of the four options represented in the above matrix, but what about the idea that memory judgements are indirectly evident judgements of fact, i.e. the idea that their evidence depends on a proof? The discussion of this possibility takes an important part of Meinong’s text, and I will be content with stressing the main articulations of his argument. To affirm that memory judgements are of this type is to say that their claim to credibility results from a general method of investigation, empirical verification. To show that such cannot be the case, Meinong proposes two arguments. The first is as follows: (1) The evidence of memory judgements is based upon a proof. (2) Being empirical, their specific proof consists in empirical verification. (3) Empirical verification is either « established by means of direct sensory perception »38 or with the help of other individuals. (4) Verifications of the first type rely on memory: the comparison with what is provided by sense-perception always depends in part on memory (to check my memory of the house by a relevant perception requires my believing that houses do not change, which in turn depends on memory). (5) Verifications of the second type have the same defect: « The memory of A is verified by the memory of B ; what significance could be attributed to the whole process, if the memory of B had not been accorded a certain degree of trustworthiness ? »39 (6) Individual verifications of memory judgements are always epistemically circular. The second is briefly sketched but can be reconstructed as follows: (1) Induction depends on gathering evidence, which constitutes the basis of inductive reasoning. (2) Most individual verifications upon which the inductive conclusion depends occurred in the past. (3) In order to be available now, they must be remembered. (4) The inductive procedure is epistemically circular. The first argument concerns any procedure of verification: it cannot be achieved without relying on the validity of memory. The second argument is about the inductive process itself, even assuming there is a way to cancel the first: memory provides the evidential basis on which the inductive procedure depends. According to Meinong, each 38 39 MEINONG, Alexius, op. cit., p. 261 (p. 198) MEINONG, Alexius, op. cit., p. 264 (p. 202) FABRICE TERONI 13 argument is decisive against the option under review, but note that there are slightly different theses in the neighbourhood, most notably: (a) It is possible to prove non-circularily the validity of a particular memory. (b) It is possible to prove the validity of memory in general. (c) It is possible to prove the validity of memory in general through the application of (a). Meinong’s first argument attempts to refute thesis (a), whereas he seems to conceive the second to be directed to thesis (b). These two questions are best kept apart, for (a) does not lead automatically to (b) : even if the validity of a particular memory is non-circularily demonstrated, this does not ground the possibility to demonstrate the validity of memory in general. The relevant possibility here is that the argument required in order to reach this further conclusion depends on memory, and thus is viciously circular. This would be the case if the only way to secure the general conclusion must rely on induction (that is, (b) can be reached only via (c)), thus falling prey to Meinong’s second argument. I think that thesis (a) can be sustained in a certain way, though not thesis (c). I cannot do justice to the complexity of this problem here, but let me briefly broach some important points. This will allow me to locate Meinong with respect to other positions. For Meinong is here in sharp opposition with a series of analytical discussions trying to demonstrate the validity of memory through sophisticated arguments40. Some attempted proofs rely on the a priori validity of the principle of induction41, which is at best problematic. But by focusing on what he calls non-retrospective cases, what I called earlier propositional memory, E.J. Furlong seems to succeed in proving the validity of some cases of memory, but is nevertheless unable to plausibly defend thesis (b): this move necessarily relies on induction, thereby succumbing to Meinong’s second argument. Note also that if what I said above on Meinong’s restrictive account of the scope of memory is along the right tracks, then the first argument goes through: there may be, as Furlong shows, non-circular ways to demonstrate the validity of the preserved belief that 2+2=4, but not judgements about what has occurred in the past made on the basis of memory presentations. Why? Because in this case, the two types of verification shown by Meinong’s first argument to be circular are the only available ones. But this is not even required for Meinong’s point to go through, for to conceive memory judgements as indirectly evident is to grasp their 40 One can cite, among others, HOLLAND, R.F., The empiricist Theory of Memory, Mind, 63, 1954 ; SAUNDERS, J.T., Scepticism and Memory, Philosophical Review, 1963 ; HARROD, R.F., Memory, Mind, 51, 1942 ; FURLONG, E.J., Memory, Mind, 57, 1948. 41 See HARROD, R.F., op. cit. 14 MEINONG ON MEMORY epistemological status as dependent upon proofs, something Meinong rightly grasps as ‘artificial’42. To opt for this conception is to sustain thesis (d) : (d) The epistemological status of any individual memory judgement depends on a proof. And there are only two options here: either to say that the justification of any individual memory judgement depends on a proof of its own validity, or on a proof of the general validity of memory. This constitutes a dilemma: on the one hand, individual memory judgements are almost never verified, and so would be unjustified; on the other, any general proof of memory depends on induction, and therefore cannot ground individual judgements non-circularily. Meinong’s most important conclusion, that memory judgements cannot be based on a proof, is thus secured. His sophisticated assessment of the epistemic circularity of any attempt to prove the validity of memory is in this way apt to cancel the attempts of many posterior analytic philosophers. But where does this leave us? Meinong has shown that the options represented in the matrix are all unfaithful to the specificity of memory judgements. There remain two options: either to admit that they are unjustified, or to extend our conception of justification in order to cover their peculiarities43. The first option is a very strong brand of scepticism, which Meinong rejects44, so what is his positive account? It is as follows: Memory judgements represent conjectures (Vermutungen).45 Conjectures are in sharp contrast with certainty, according to Meinong, and this difference can be described in terms of psychology as a difference in the intensity of the act of judgement.46 Psychologically speaking, the difference between memory and introspective judgements consists in distinct degrees of confidence: the subject simply feels more confident in the later than in the former. Meinong further notes that with conjectures we face characteristics immanent in the types of judgement concerned.47 42 MEINONG, Alexius, op. cit., p. 265 (p. 203) Note that, for reasons similar to those given by Meinong, perceptual judgements cannot either be classified in the matrix. On this point, see CHISHOLM, Roderick, Verstehen : The epistemological Question, in Foundations of Knowing, Minneapolis, 1982. 44 MEINONG, Alexius, op. cit., p. 265 (p. 204) 45 MEINONG, Alexius, op. cit., p. 266 (p. 204) 46 MEINONG, Alexius, op. cit., p. 266 (p. 205) 47 MEINONG, Alexius, op. cit., p. 267 (p. 206) 43 FABRICE TERONI 15 As I understand him, this means that for conjectural judgements, though not for directly evident ones, we can detach justification from factuality: a conjectural judgement which comes out false can still be justified, whereas a false directly evident judgement is an impossibility. In a nutshell, conjectural judgements are not as immediately connected as evident ones to the way the world is (or was). Memory judgements have direct conjectural evidence. Because they cannot be epistemically grounded on proofs, their justification must be intimately linked with the judgement itself and hence direct; because they are distinct from evident judgements, they are endowed with a specific kind of justification, conjectural evidence, which is compatible with error. Meinong concludes that every individual judgement carries its whole guarantee in itself (…).48 A memory judgement is by itself, because it is a judgement of this type, endowed with its own specific (and faillible) guarantee. This is not the case for every kind of judgement: a mathematical judgement, for instance, is not justified for the simple reason that it is so. The specificity of memory judgements, as we saw in section (1), is to be made on the basis of presentations: this should motivate a specific account of memory presentations, thereby explaining their epistemological status49. Memory judgements are made on specific bases, and are eo ipso endowed with justification. In this section, I presented Meinong’s way to ground his position on the epistemological status of memory judgements. The conclusion is reached via a process of elimination. This very interesting procedure has had a fundamental impact on Roderick Chisholm50, and grounds a sophisticated form of foundationalism that does not look for unshakable foundations in any area of knowledge. Meinong is here in profound agreement with many contemporary foundationalists whose ancestor is Thomas Reid51. The relation between these two thinkers could constitute the topic of another discussion, so let me simply note that they share a form of faillibilism, the idea that justified beliefs can be wrong, and of particularism: some individual judgements possess their own guarantee. Moreover, Meinong’s discussion is closely related to contemporary debates on the distinction between being justified and proving that one is, which William Alston applies against many contemporary epistemologies52. Meinong, by refusing thesis (d), 48 MEINONG, Alexius, op. cit., p. 268 (p. 207) Something Meinong does not provide in his essay. 50 See CHISHOLM, Roderick, Foundations of Knowing, Minneapolis, 1982, and Theory of Knowledge, Prentice-Hall, 1989 51 REID, Thomas, Philosophical Works, Olms Verlag, 1983 52 See ALSTON, William, Epistemic Justification, Cornell UP, 1989 49 16 MEINONG ON MEMORY and by stressing issues of epistemic circularity, displays a profound sensitivity to this fundamental distinction. (3) BRENTANO’S CRITICISMS AND THE NATURE OF PRESUMPTIVE EVIDENCE Before concluding, I want to briefly discuss the way we have to understand Meinong’s position, which can be done by discussing Brentano’s criticism. Let me first mention some points about Brentano’s conception of evidence. He contrasts what he calls blind with evident judgements53. When evident, a judgement is incompatible with error, as well as doubt54. Brentano therefore works with a restricted account of evidence, which corresponds to what Meinong calls direct evidence. He moreover distinguishes between mediately and immediately evident judgements, something we have already met above and which corresponds to the presence or absence of proof. Armed with this conception of evidence, Brentano addresses a fundamental criticism to Meinong’s position based on the nature of knowledge of probabilities. He points first that we can be certain of probabilities, for instance of the fact that the probability of throwing this dice and scoring a six is 1/655. This kind of judgements can be evident, but what the subject judges with evidence is the holding of a probability. What is, more generally, the structure of this kind of knowledge? In reference to Laplace, Brentano notes that each probability is composed of knowledge and ignorance, of which we must be aware56. Thus, for instance, when Sam judges that it will probably rain tomorrow, his judgement is constitutively linked with his awareness that, let us say, the present weather is a sign of rain, and with his awareness that it may not rain. The presence of knowledge and ignorance explains why he makes a probabilistic judgement. This immediately grounds this objection to Meinong: because probabilistic judgements are composed of two elements, the notion of immediate presumptive evidence is contradictory. Such judgements are essentially mediately grounded, because their evidence depends on the weighting of reasons. Meinong thus faces a dilemma: either to show that memory judgements are not probabilistic, or to renounce his appealing position on the problem of memory. Note first that we should incline to interpret him as refusing to identify memory judgements with judgements of probability understood in 53 See for instance BRENTANO, Franz, Versuch über die Erkenntnis, Felix Meiner Verlag, 1970, p. 178 54 BRENTANO, Franz, Wahreit und Evidenz, Felix Meiner Verlag, 1974, p. 144 55 BRENTANO, Franz, idem, p. 145 56 BRENTANO, Franz, idem, p. 145 FABRICE TERONI 17 Brentano’s way, because to gather evidence for or against the occurrence of an event probably depends epistemically on memory (and presumably also on judgements about remembering itself), and we saw how sensitive Meinong is to these problems. Moreover, he never uses knowledge of probabilities in order to illuminate his account of memory, but in this case memory judgements would be « based on something which [they] have in common with others that epistemologists have already investigated »57, something he never points out. Nevertheless, there is a passage translated in a way that sustains Brentano’s reading: If a person (…) increases the intensity of his memory judgement to certainty, then the failure of verification will obviously reveal his error ; but is he now going to renounce his memory judgement in the same way as he would have to cease to trust a mathematical axiom at any time if this axiom had (…) been included in a calculation and led to a wrong result ? By no means ; his exaggerated confidence is somewhat reduced but not destroyed, and no one has succeeded in exorcizing it. »58 Let me first construct this passage in what I take to be Brentano’s way, before suggesting another interpretation. For it is indeed puzzling. In the first sentence, is it implied that if the judgement is not made with certainty, then the failure of verification will not reveal his error? By MTT, we have at least: if the failure of verification does not reveal his error, then the judgement was not made with certainty. And Meinong seems to imply that this is the case for memory judgements, which are compatible with error. But the compatibility of a judgement with the non-factuality of its objective occurs only for probabilistic judgements. Hence, via Brentano’s plausible remarks, we reach a dramatic conclusion. This is unsatisfying on two counts. First, in the case of probabilistic judgements, the degree of conviction must be tuned to the evidence gathered for and against the occurrence of an event, and this would, as noted above, run against the whole thrust of Meinong’s argument. Second, this means that memory judgements are probabilistic: for instance that it is more probable than not that I had some specific mental state in the past. This does not respect the nature of memory judgements in two ways. (a) The suggestion is psychologically implausible: normal memory judgements simply do not have this form. And (b), dynamic considerations show that this is wrong: when I make a memory judgement about Mary’s party, and later find a reason against its having taken place, I do not modify the strength of my inclination to believe that such was the case, but more dramatically cancel the judgement. Moreover, if the judgement is probabilistic, then I am not required to modify its intensity even in the face of ‘error’. For judging 57 MEINONG, Alexius, op. cit., p. 256 (p. 192), see the beginning of section (2) above. 58 MEINONG, Alexius, op. cit., p. 268 (pp. 207-8) 18 MEINONG ON MEMORY with certainty that the probability of a specific past party having occurred is 75/100 is compatible with its not having occurred. More generally, two things should not be confused: the objective of the judgement and its intensity. Any objective is compatible with any grounded intensity: to judge that the probability of throwing a dice and scoring six is 1/6 with certainty is sound in some gambling joints, as well as to judge with presumptive evidence that there was a party. Meinong’s passage can be reconciled with these criticisms by looking more closely at the original: the translation is misleading because it uses ‘memory judgement’ twice, whereas Meinong writes, first, ‘die Intensität seiner Gedächtnisurteile’ (plural) and next ‘das Urteil seines Gedächtnisses’. Thus, he can be interpreted as saying that someone who is certain of his memory judgements in general will not renounce to trust memory because one of them is mistaken, but will tune his trust to their specific evidence. This does not imply that particular memory judgements can be maintained in the face of defeating evidence, and hence they need not be understood as probabilistic. So, if Meinong does not defend the implausible position Brentano attributes to him, what is his? The most appealing position along Meinongian lines make use of the distinctions just mentioned in the following way. First, by arguing that memory judgements are not, for the reasons sketched above, probabilistic: to judge on the basis of memory that there was a fire in one’s house is not to judge that this event probably occurred. Second, by stressing something I mentioned before but neglected since then: presumptive evidence should be given a normative reading. Accordingly, to make a memory judgement is to have the right to make a judgement, and its intensity has to match the specific evidence. This evidence is moreover presumptive: this means, according to Meinong, that one has no right to be certain (and is usually not). But this is not to say that we should conceive memory judgements as probabilistic, for we can give a more appealing account of the way subjects can be tuned to the specific evidence of memory: by being open to error. This is to stress that presumptive evidence is a way judgements are made, and not judgements with specific objectives. The idea is therefore that when remembering, but not when one has direct evidence, one should be open to correction. Someone who systematically maintain his memory beliefs on the face of overwhelming evidence to the contrary is in this sense irrational. But to say that subjects must be open to correction is not to say that their judgements are probabilistic: they are not, but subjects are aware that if they encounter some specific kinds of defeating evidence, they should abandon them. We reach a foundationalist position according to which judgements based on the deliverances of our primitive faculties are justified until proven guilty, in Thomas Reid’s words, or FABRICE TERONI 19 that they are prima facie justified59, according to the current idiom. Prima facie justified judgements are not evaluated as unjustified when defeating evidence is found, thus their epistemic status is compatible with error, but are not maintained, with any degree of confidence, once counterevidence is present. Thus, it is possible to remember without believing, for instance when one mistakes misinformation as defeating evidence, as well as being justified and wrong. (4) CONCLUSION Meinong’s essay on memory constitutes a fundamental discussion of this basic endowment of the mind. I argued above that his conception of memory, as well as his remarks on different memory judgements are important in the context of contemporary debates. Meinong’s own brand of memory foundationalism is also very interesting. It appeals to internal bases of judgements: a memory judgement is justified because it occurs in the context of a memory presentation relevant to its subject-matter. This is in sharp opposition to contemporary forms of reliabilism, and I would add far more plausible. His particularism, his sensitivity to issues of epistemic circularity, as well as his faillibilism, ground fascinating positions on the epistemic structure of fundamental judgements. Many contemporary epistemologists have been seduced by these antidotes to classical foundationalism, and can only profit from interaction with Meinong’s seminal discussion of memory. 59 On prima facie justification, see, among others, POLLOCK, John, Knowledge and Justification, Princeton, 1974 ; ALSTON, William, Epistemic Justification, Cornell UP, 1989 ; SOSA, Ernest, Knowledge in Perspective, CUP, 1991 ; and CHISHOLM, Roderick, Theory of Knowledge, Prentice Hall, 1989