Marty on Truthmaking

To appear in: L. Cesalli / J. Friedrich (Eds.), Anton Marty and Karl Bühler. Between Mind and Language, Schwabe 2013. Marty on Truth-making Arkadiusz Chrudzimski* (Szczecin) In Marty’s writings one can find two theories of truth-making. The first one is deeply rooted in Brentano’s picture of intentionality as put forward in his unpublished lectures and operates with both immanent and transcendent propositional entities. Marty’s second theory was formulated under the influence of Brentano’s late criticism of his own theory of immanent objects. It rejects all immanent entities, but differs significantly from Brentano’s late, “reist” position. I begin this paper with a few words about the general idea of truth-making. Then, in section 2, I sketch Brentano’s official theory of intentionality and describe a simple theory of truth-making that can be based on it. Section 3 contains an analysis of Marty’s first theory, employing the ontology of intentionality from Brentano’s unpublished lectures, and section 4 concerns Marty’s late theory, in which he rejected all immanent entities, but retained a rich ontology of external states of affairs. 1. Truth and truth-making We all know that such things as news, statements, or beliefs can be true or false. Moreover their truth or falsity is a rather important feature. In fact we invest a significant portion of our energy into pursuing truth and avoiding falsehood. Therefore it would be nice to know what truth is. This sounds like a deep philosophical question, but, surprisingly enough, our common sense gives us a quite precise idea of truth. “Things” that can be true or false seem always to be “about something” and they are true just in case this “something” is the way they say it is. This is, in a nutshell, the layman’s idea of “corresponding with the facts” and the eternal source of philosophers’ “realist” or “classical” intuitions about truth. Now this rudimentary construal of truth has been criticised in many ways. On the one hand, for some thinkers, it appears trivial and philosophically uninteresting (like Putnam * The work on this paper was supported by the Austrian Foundation for the Promotion of Scientific Research (FWF) and the Foundation for Polish Science (FNP, “Master” programme scheme, directed by Tadeusz Szubka). 1 1995); on the other hand, others argue that, far from being trivial and self-evident, it is rather irreparably paradoxical and finally self-destructive.1 Only after Tarski (Tarski 1933) did we learn that a significantly refined version of the classical concept of truth can be consistently applied, although at the expense of a transformation of our seemingly universal discourse into an open-ended hierarchy of meta-languages. No doubt this is a high price and it is therefore no surprise that there are still many influential philosophers who try to persuade us that the classical concept of truth is at best empty and philosophically uninteresting, and at worst, deeply inconsistent. (cf. e.g. Putnam 1995; Dummett 1976, 116) The critics of the classical construal of truth typically begin with the empirical claim to the effect that every concept has to involve some criteria of application (cf. e.g. Putnam 1995, 10; Dummett 1976, 116) and consequently propose some versions of its epistemologization. Roughly a realist conception of truth assumes that truth consists in a certain relation of a truth-bearer to the pieces of reality which it is “about”, while an epistemic definition claims that truth is to be defined in epistemic terms like evidence, coherence, or (idealized) rational acceptability.2 Such a distinguished epistemic status may be an intrinsic property of a truthbearer (like Cartesian “clearness and distinctness”) or it may consist in some relation, but this time it is not the relation that the truth-bearer bears to the reality which it is “about”, but rather its relation to some other truth-bearers.3 There are also deflationist approaches denying that the word “truth” refers to any substantial property at all. 4 (Cf. Alston 1995, 41f.) Now, in spite of the fact that the issue of the realist concept of truth is far from being settled, in the following I am going to assume that it is a sensible philosophical option. At the very least it was Marty’s opinion and in this paper I am going to analyse his theory of truthmaking.5 1 This attitude was characteristic of the philosophers belonging to the early logical positivism. Precisely for this reason both Carnap in his Logische Syntax der Sprache (Carnap 1934) and Popper in his Logik der Forschung (Popper 1935) strictly avoided operating with the concept of truth. 2 Cf. “[T]ruth is an idealization of rational acceptability. We speak as if there were such things as epistemically ideal conditions, and we call a statement ‘true’ if it would be justified under such conditions.”, Putnam 1981, 55. Cf. also and Sellars 1967, S. 140–142. 3 Typically it will be coherence with other beliefs of a given subject, derivability from some “basic beliefs”, correspondence with what the members of an ideal scientific society would believe “in the long run” (Peirce 1868, 52, Peirce 1878, 139), correspondence with what is believed by God, and so on. 4 Locus classicus is Ramsey 1927/1990. 5 By the way, Marty’s realist construal of truth shouldn’t be interpreted as an unreflective acceptance of the “default” theory. He developed his views in direct opposition to Brentano who defended an epistemic definition of truth. 2 The idea of truth-making as developed in the last few decades6 is a certain version of the classical understanding of truth. According to this understanding, the truth of a truthbearer depends on the way the world is. The idea of truth-making in its rudimentary form says that at least for certain true truth-bearers there is something in the world that makes them true. Spelling out this idea is not an easy task and I won’t try to do this within the limited scope of this paper. So let me only signal some important points. (i) First of all there is an issue of analytic truths that are often interpreted as “vacuously true” by people impressed by the early Wittgenstein. Indeed it is not implausible that they would be true even in an “empty universe”, i.e. without anything that can function as truth-maker. (ii) Second, even in restriction to the non-analytic truth-bearers, there needn’t be any one-to-one correlation between truth-bearers and truth-makers. At least for some important kinds of composition, the truth values of compound sentences will be obviously parasitic on the truth values of its components. Consider the familiar rules for the extensional logical connectives: “Non-A” is true iff “A” is not true “A and B” is true iff “A” is true and “B” is true “A or B” is true iff “A” is true or “B” is true “If A then B” is true iff “A” is false or “B” is true “A if and only if B” is true iff “A” and “B” are both true or both false Applied to the truth-making theory, they mean that we need no distinct truth-makers for compound sentences that are obtained by means of the listed extensional connectives. If there exists a truth-maker for the sentence “A”, then the sentence “A or B” has also been made true; if we have truth-makers both for “A” and for “B”, then the sentence “A and B” has also been made true; and all that must be done in order to make the sentence “Non-A” true is that there is no truth-maker for the sentence “A”. If we follow this idea, we need no conjunctive, disjunctive or negative facts in the world. In general it is to be expected that we will need truth-makers only for our true atomic truth-bearers.7 6 See Mulligan/Simons/Smith 1984. Armstrong (2004, 1) claims that the most important motivations for the idea of truth-making came from C. B. Martin and his discussions of phenomenalism. 7 But remember that many connectives of the natural language do not seem to be extensional. In particular the idea, that all philosophically important “if-then” connections can be analysed in terms of the material implication 3 (iii) Third, it seems that the relation between truth-bearers and “their” truth-makers that the truth-making theorists have in mind is to be understood rather as a necessary connexion, than as a merely contingent coincidence. It could happen that every truth has its admirer. If it were the case, then there would be, for every true truth-bearer, a subject who admires it. But even if it were so, it would be a merely contingent correlation. Nothing in the concept of a true truth-bearer forces us to assume that there is somebody who admires it. Now the relation of truth-making seems to of a different kind. The fact that every (with restrictions of the kinds (i) and (ii)) true truth-bearer has its truth-maker seems to be encapsulated in the very concept of truth as understood by the truth-making theorists. Therefore the general theory of truth-making will have to be formulated in modal terms. (iv) But even assuming the constraints (i)–(iii) we still get a very general formulation which allows almost anything as a truth-maker. In particular it even doesn’t exclude epistemic theories of truth, if we allow the property of having a certain epistemic status as an entity that can play the role of truth-maker. So if we want to formulate a narrower concept of truthmaking which will correspond more to the realist intuitions, we should restrict the scope of possible truth-makers to those pieces of reality which the relevant truth-bearers are “about”. It is not an easy task to formulate this condition in a general form, i.e. without assuming a particular semantics (or a particular theory of intentionality) explaining the mechanism of “being about”. Concerning this last point I hasten to add that in this paper I am going to operate within the framework of particular theories of intentionality that explain quite precisely this mechanism. It is to be expected that such a theory will make explicit three following concepts: (A) truth-bearer; (B) truth-maker; and (C) the relation of “making-true” between them. (A) The question about truth-bearers is the question of the category of entities that are primarily capable of being true or false. Many analytically trained philosophers would without hesitation point at sentences, those acquainted with the work of Austin and Searle possibly at speech acts. In a sense they are perfectly right. Of course we often ascribe truth and falsity to our uttered sentences, but the big question is whether sentences or speech acts are primary truth-bearers. Behaviourists in the style of Quine and Sellars, and maybe some champions of causal semantics in its radical reading, would say “yes”, but a more mundane view, shared by the way by such distinguished thinkers as Chisholm and Searle, has it that our speech acts can is just a fantasy from the heroic days of analytic philosophy. It is therefore quite probable that we will need some special truth-makers for modal and causal conditionals. 4 be truth-bearers only because they express our mental states. The mental states, on the other hand, do not express anything. They are primary truth-bearers that have intrinsic semantic (or intentional) properties. This was the view of Marty and the whole Brentano School. In the following I am going to assume this kind of “mentalism” and claim that the primary truthbearers are mental states traditionally called “beliefs” or “judgements”. (B) Primary truth-bearers have intrinsic intentionality and if they succeed in hitting the relevant parts of reality, then these parts of reality can be said to function as their truthmakers. Generally one can distinguish between two approaches to the ontology of truthmakers: (a) the doctrine of an unarticulated truth-maker to the effect that the world as a whole is the only “big” truth-maker (Frege’s doctrine of “The True” and “The False” can be interpreted this way); (b) theories of an articulated truth-maker assuming that different parts (or aspects) of the reality are responsible for truth of different truth-bearers. Now concerning (b) the most important ontological question pertains to the nature of these parts. From the history of philosophy and contemporary discussions, we have various proposals. After Wittgenstein, a standard move is to speak of states of affairs. Some more parsimonious philosophers propose tropes (Mulligan/Simons/Smith 1984). Even nominal objects can play this role, if we assume a certain analysis of truth-bearers (see section 2 of this paper). Particularly the truth-makers of the first and last category will be important for the further discussion. (C) Finally we have the relation of making-true. It obtains between the true truthbearers and “their” truth-makers. Its “material character” will vary depending on a given theory. But there is a certain important formal condition that, as it seems, must be met by any theory of truth-making. The relation of making-true has to be an internal relation in Russell’s sense, which means that it is necessarily generated by monadic natures of its terms. In general, if relation R is internal, then the following holds: necessarily, for every x and y, Rxy, if and only if there are some monadic properties φ and ψ of a and b which are such that, for every pair of objects z and w, if z has φ and w has ψ, then Rzw. (Cf. Russell 1910, 374; Johansson 1989, 119f.) In this sense, internal relations supervene on the monadic natures of their terms. The main concern of an articulated theory of truth-making is to explain how exactly the relation of making-true supervenes on its terms. Typically this explanation will refer to the structural features of truth-bearers and truth-makers.8 8 A paradigm example is of course Wittgenstein’s theory in the Tractatus to the effect that the relation of making-true is nothing but the identity of configuration (the configuration of simple names in a sentence and the 5 2. The general Brentanian Framework Let us now move to Marty. In his works he formulated two theories of truth-making. Particularly the first of them depends heavily on Brentano’s theory of intentionality. So let me sketch in this section the general Brentanian picture. 2.1 Intentional objects Intentional directedness is explained by Brentano in terms of intentional objects (Brentano called them “immanent objects”). The most basic intentional state, the starting point of any intentionality, is called presentation (Vorstellung). According to Brentano it consists in having an immanent object before one’s mind. It makes no difference whether the “external” object of a presentation exists or not. The immanent one exists always. Without it, a presentation simply wouldn’t be an intentional state. (Brentano 1874/2008, 106). There are two possible readings of this theory: an ontologically neutral, and an ontologically robust one, depending on how seriously Brentano’s immanent objects are taken. It seems that the ontologically neutral understanding of immanent objects, resembling the medieval doctrine of esse obiectivum, prevailed in Brentano’s early writings, while the ontologically robust interpretation emerged in the middle period of his philosophy.9 Roughly, on the ontologically neutral reading, the discourse introducing immanent objects is to be interpreted as an ontologically non-committing façon de parler, while on the ontologically robust construal, intentional objects are to be taken seriously as an important ontological category. Happily I need not go into details of this controversy here, because only the ontologically robust reading is relevant for Marty’s theory. Marty states explicitly that it makes no sense to ascribe to immanent objects any special (in particular any “ontologically weaker”) mode of being. True enough immanent objects belong to a different ontological category than external things, but they have exactly the same mode of existence.10 In general Marty is, contrary to Aristotle, a partisan of univocity of “being”. configuration of simple objects in a state of affairs). 9 The ontologically robust version of the theory of immanent objects can be found in Brentano 1891/1980. Cf. also Chrudzimski 2004, chapter 3. For Brentano’s early theory see Chrudzimski 2004, chapter 2. 10 Cf. „Die Scholastiker unterschieden eine mentale und reale Existenz. Mental oder, wie man sich auch ausdrückte, ‘objektiv’ (d.h. als Objekt einer psychischen Tätigkeit) existiert ein Vorgestelltes, Geliebtes als solches. Real existiert der Vorstellungsakt, eine Ausdehnung u. dgl. Damit können aber offenbar nicht zwei verschiedene Bedeutungen der Existenz gemeint sein. Der Begriff, ‘die Anerkennung zu verdienen’, ist stets derselbe, mag er wem immer, einem Realen oder Nichtrealen (z.B. speziell einem ‘Objektiven’ als solchen) zukommen. Nur das soll eben gesagt sein, daß, wer das eine anerkennt, damit ein Reales anerkannt habe, wer das 6 Here is a general picture of presentation as construed by an “ontologically serious” Brentanian: ϕ identitfying property intentional object has the identitfying property “in a sense” (Zalta’s encoding) the reference object has the identifying property “in the standard sense” (exemplification) intentional object IMM REPR intentional reference subject reference object The conscious subject on the left refers intentionally to the object on the right. He does it “through” the property ϕ. The reference-object “is meant” by the subject as a ϕ-object (say as a Christmas tree). This is of course the (in)famous identifying description theory of intentionality in it’s Brentanian version. According to this view an intentional reference is only possible by “listing” some identifying properties of the target object. Therefore in the following I will refer to the property ϕ as the identifying property. The required “listing” of identifying properties is in turn only possible because the subject’s mind is able to produce an immanent object that “has” the required properties (in our case: the property of being a Christmas tree). In our picture, this object is placed in the comic-style “cloud” to illustrate the metaphor of being “in the mind”. andere, ein Objektives als solches. Reale Existenz heißt also Existenz eines Realen. Das Adjektiv ist ähnlich verwendet, wie wenn man anstatt von der Bejahung eines allgemeinen Inhalts, von ‘einer allgemeinen Bejahung’ spricht.”, Marty 1884, 173f. 7 However the kind of having the property ϕ by the immanent object is a bit special. It would surely be strange to think that a mind-dependent immanent object can be a Christmas tree in the literal sense; and indeed, Brentano says that immanent objects have their identifying properties only in a “non-genuine”, “modified” sense. (Brentano 1891/1982, 26f.) Following Zalta I will call this kind of having a property “encoding”. (Zalta 1988, 16f.) The intentional reference is successful if and only if there is in the world an object that exemplifies all properties that are encoded by the intentional object. Two important relations are involved in this picture – the relation between the subject and the intentional object, and that between the intentional object and the reference object – are called IMM and REPR respectively. These names suggest “immanence” and “representation”. IMM is construed as constitutive to any intentional reference while REPR obtains only if the reference is successful.11 Again following Zalta, let me symbolise “a exemplifies F” as “Fa” and “a encodes F” as “aF”. Assuming this notation (and a quantification over predicate variables) one may formulate the following condition for obtaining of REPR: ∀x∀y[REPRxy ≡ ∀φ (xφ ⊃ φ y)]. The relation of representation obtains between x and y (x represents y) if and only if each property that is encoded by x is exemplified by y. Now the definition in this form has a certain strange consequence. According to it, an entity that doesn’t encode anything (i.e. every entity that isn’t an intentional object, like Alpha Centauri, Mt. Everest or Bill Clinton) would represent everything. To avoid this consequence, we can add a condition to the effect that the first term of the relation REPR has to encode something (i.e. has to be an intentional object): ∀x∀y[REPRxy ≡ ∃ψ ( xψ) ∧ ∀φ (xφ ⊃ φ y)]. 11 For the readers interested in general ontological issues let me stress that according to Brentano and his students, the properties involved are not universals but tropes. The kind of symbolizing used in my diagrams can suggest that they are universals literally shared by numerically different individuals, but this is only a question of depiction. It would be perfectly possible to use two symbols for two individual properties (say ϕ1 and ϕ2) and connect them by the relation of strict similarity. I omit this complication for the sake of simplicity. 8 As we see, REPR is an internal relation. It supervenes on the monadic natures of its terms. What is special about REPR compared with such supervenient relations as being taller than or being of the same colour as is the fact that the supervenience basis on the side of the immanent object contains not its exemplified but its encoded properties. In fact the properties genuinely exemplified by the intentional object (like e.g. being ontologically dependent on a conscious subject) are absolutely irrelevant for the singling out of the target object and the constitution of REPR. 2.2 A non-propositional theory of judgement In Brentano’s published works this theory of rudimentary mental directedness is supplemented by a highly interesting, non-propositional theory of judgement. All judgements are interpreted as existential acceptances or rejections of nominal objects, according to the following scheme: intentional object has the identitfying property “in a sense” (Zalta’s encoding) Judgement + / – ϕ identitfying property the reference object has the identifying property “in the standard sense” (exemplification) immanent object REPR Presentation intentional reference subject reference object We see that this new picture doesn’t change much. In addition to what we have already seen we have only a new intentional relation, labelled as “Judgement +/–”, which is to be 9 interpreted as a mental attitude of “acceptance” or “rejection” of the presented object. So according to Brentano a judgement always assumes a presentation (is ontologically founded on it) and adds to it only a new mental mode, but no new objects or contents. The claim that the object of a judgement is exactly the same as the object of the underlying presentation, was one of the core theses of his official theory of intentionality. (Brentano 1874/2008, 231)12 From the point of view of ontological parsimony it is without doubt a very attractive theory as it introduces no special propositional objects, but a big challenge is of course to show how all the various judgement forms can be analysed in terms of existential acceptances and rejections. Brentano tried to show this. Simple existential judgements like “Santa Claus exists / doesn’t exist” pose no problem for his theory, and for the other judgements of apparently predicative form he proposed an interesting reformulation. If we symbolise the 12 There is an aspect of the above diagram that can strike us as very counter-intutive. I mean the fact that the judgemental acceptance or rejection is directed at the immanent object and not at its transcendent counterpart. What we would expect, is rather that this mental attitude will be “about” the genuine (transcendent) reference object. After all the immanent object exists always, and therefore every rejecting judgement would automatically turn out false. The issue is in fact very perplexed. First of all one has to remember that Brentano’s non-propositional theory of judgement was developed in the early period of his thought. At this time Brentano’s theory of intentionality used the rhetoric of the medieval doctrine of the ens obiectivum. Within the framework of this doctrine an immanent object is interpreted as something which has a peculiar (and probably ontologically noncommitting) mode of existence “in the subject’s mind”, but which can also exist “outside the mind”. Furthermore there is no mention of the modified sense in which immanent objects have their identifying properties. Under these assumptions it is much easier (and indeed quite natural) to assume that it is precisely the immanent object that is accepted or rejected in a judgement. But later this non-propositional theory of judgement has been directly implemented in the theories of the middle period where immanent objects are reinterpreted as ontologically serious entities with a very peculiar internal structure. In particular they have their identifying properties in a modified sense, so that it is no longer natural and unproblematic to say that a judgement is directed at an immanent object. On the other hand it must be stressed that Brentano has in fact no other choice. An important principle, clearly formulated in his middle period, but present also in his early philosophy, was that every intentional attitude involves a genuine relation to “its” object (see Chrudzimski 2001, 20f.). The consequence is that only immanent objects can play the role of objects of mental acts. All this doesn’t looks very promising and Brentano’s own struggle with this predicament wasn’t in the end successful. As we know he finally banned immanent objects from his late ontology as useless and fictitious entities. I believe that Brentano’s middle theory of intentionality in fact involves many fundamental difficulties but in this paper I cannot go into details (see Chrudzimski 2001, 80–83). They will be however not so important for Marty’s theory, because Marty replaced Brentano’s official non-propositional approach by a kind of propositional theory of judgement where the mentioned difficulties don’t appear directly. 10 existential acceptance as “+” and the existential rejection as “–”, Brentano’s translation goes as follows: A is B – + A which is B A is not B – + A which is non-B Brentano provides an existential translation also for the four traditional Aristotelian forms (A, I, E, O). (Brentano 1874/2008, 236). If we symbolise the existential accepting / rejecting as “+/–”, use a negation operator “*”, allowing us to build a negative counterpart of any given term, so that “*A” means “non-A”, and symbolise the connection “which is” between properties simply by the concatenation of terms, then the four Aristotelian forms will look like that: AaB (Every A is B) – – A*B AiB (Some A are B) – + AB AeB (Every A is non-B) – – AB AoB (Some A are non-B) – + A*B We assume here that the operator “*” has always the minimal scope, so that “*AB” means “a non-A which is B” and not “a non-A which is non-B”. This is without doubt an excellent analytic work from the borderline of the philosophical logic and psychology, promising a significant ontological simplification, but it has its price. If Brentano’s existential reduction is to work at all, then all the structure that we typically put into propositional contents has to be transferred into nominal objects. Two problems emerge immediately. First of all we have the connection of properties (symbolised above by the concatenation of terms). Second, we have strange negative properties. Both problems have been pointed out by Meinong (Meinong 1892, 205f) and they led him eventually to a quite complex theory of propositional entities. (Meinong 1902, cf. also Chrudzimski 2009). The further problems (also crucial for Meinong) arise as soon as we take relations seriously. It is generally far from clear how they are to be handled within the framework of Brentanian philosophy.13 13 Brentanians tended generally to construe all relations as internal. If this position were tenable, it would of course marginalize this problem. 11 2.3 Brentanian truth-making I believe that the above mentioned problems are in fact fatal for the project of Brentano’s existential reduction, but in this paper, I am interested only in analysing Marty’s theory of truth-making. Now, since Marty accepted Brentano’s theory of judgement wholeheartedly, let us also pretend that it works. So, if we accept this picture of judgement, we are in a position to construct a very simple and elegant theory of truth-making. It goes as follows: a positive judgement of the form “+ A” is true if and only if the involved representation in fact represents something (i.e. if there is (an) A); a negative judgement of the form “– A” is true if and only if the involved representation does not represent anything (i.e. if there is no A); and as we remember the relation of representation has been defined as follows: ∀x∀y[REPRxy ≡ ∃ψ ( xψ) ∧ ∀φ (xφ ⊃ φ y)] According to this proposal, all the truth-makers we need are objects of the nominal form (although sometimes of a quite complicated internal structure). An object a makes true a positive judgement b if and only if a exemplifies all the properties that are encoded by the immanent object of the presentation involved in the judgement b; and a negative judgement c is true if and only if there is no object exemplifying all the properties that are encoded by the immanent object of the presentation involved in the judgement c. This theory of truth-making contains an interesting asymmetry. What makes a true negative judgement true is not a special entity but rather the absence of an entity that would make the correlative positive judgement true. And it must be stressed that this absence is not to be interpreted as an extra entity, but rather as a “pure lack” without any ontological interpretation. Soon we will indeed have to talk about such entities as “non-existence of A”, but within the framework of this simple picture we don’t need them in our ontological universe. Let me close this section with a word of warning. What has been presented here as a Brentanian theory of truth-making shouldn’t be interpreted as Brentano’s official theory of truth. His official construal of truth was an epistemic theory to the effect that a judgement is true if and only if it could be made by someone who judges with evidence.14 Brentano saw no possibility for a consistent formulation of a realist theory of truth for the reasons that were 14 Sometimes it is claimed that the epistemologization of the concept of truth was characteristic only of the late Brentano (cf. e.g. Srzednicki 1965), but in fact the epistemic construal of truth prevails in all periods of his philosophy. Cf. Chrudzimski 2001, chapter 2. 12 based on his concept empiricism. (Cf. Chrudzimski 2001, 71ff.) So the theory of truth-making presented in this section can be at best called “pseudo-Brentanian”. It has been formulated here for two reasons. First of all, it seems to be interesting and attractive on its own; and second, it will be helpful for the further analysis of Marty’s theory. Marty’s construal of truth was always realist and he believed that this position can be perfectly combined with the doctrine of concept empiricism. (Cf. Chrudzimski 2005a, 84f.) 3. Marty’s early theory Marty’s philosophy can be divided into two big periods. In this section I’m going to focus on his early theory (at least until 1904). At this time Marty accepted the Brentanian theory of intentionality (in its ontologically robust form), but additionally introduced both immanent and transcendent states of affairs that were called in the Brentano-school “contents” (Inhalte). (Cf. Marty 1884, 301) 3.1 Brentanian theory of propositional contents In the light of my presentation of Brentano’s official views in section 2 it may sound strange, but in fact Marty took his account of propositional contents from Brentano. True enough, they were absent in Brentano’s published writings, but they can be found in his unpublished lectures.15 The theory of propositional contents that was developed in the Brentano school was a rather strange construction. It shouldn’t be understood as an ontological answer to the fundamental, Meinong-style criticism referred to above. This criticism tries to introduce states of affairs by showing that Brentano’s existential reduction fails, that his “nominal” objects are in fact states of affairs in disguise. The theory of propositional contents I am going to describe now ignores this criticism and assumes the success of Brentano’s existential reduction! From the start it is taken for granted that all the states of affairs have existential form corresponding exactly to the possible structures of Brentano’s judgements. In the previous section we have seen that under this assumption we are in principle able to build a functioning theory of intentionality with the help of nominal objects alone and from this perspective the Brentanian 15 Cf. above all Brentano’s Logic Lectures (Manuscript EL 80). On the theory of intentionality developed in these lectures see Chrudzimski 2001, 42–46, 62–66. In his later writings Marty acknowledges his debts to Brentano. Cf. “So habe ich selbst […] den Terminus ‘Urteilsinhalt’ verwendet und vor mir Brentano in seiner Würzburger und Wiener Vorlesungen.” (Marty 1908, 292) 13 theory of propositional contents looks like a pure ontological inflation. So why did Brentano introduce them? His reasons seem to have to do with the “descriptive” character of his philosophy. Brentano noticed that we sometimes seem to refer to such contents. It happens in sentences like “It’s a fact that there are no centaurs”, which in Brentano’s translation would read: “The non-existence of a centaur exists”. Instead of trying to explain this phenomenon away, he took it at face value and incorporated an ontologically robust theory of propositional contents. In his lectures Brentano introduced both immanent and transcendent contents. An existential judgement, consisting in acceptance or rejection of an object A, has now as its immanent correlate an immanent content, which Brentano sometimes calls “accepted/rejected A”. If the judgement is true, then it corresponds to a transcendent content which is called “existence/non-existence of A”. In the case of a true positive judgement it looks like this: E! encoding ϕ encoding existenceproperty identitfying property exemplification immanent content Judgement + exemplification immanent object Presentation REPR intentional reference transcendent content reference object subject In comparison with the previous diagram the picture has changed significantly. First of all, the existential acceptance now has certain ontological consequences on the immanent side. It is 14 not only a new mental mode (of accepting or rejecting), but it produces also its own immanent correlate – a new immanent entity which in a sense “contains” the object of the original presentation. Second, provided the judgement in question is correct, there must be on the side of transcendent entities also a corresponding propositional entity. It also seems that this doctrine may be interpreted in terms of Zalta’s distinction between exemplification and encoding, if we construe the immanent content as encoding the property of existence (symbolized here as “E!”) by the immanent object of the underlying presentation, and the corresponding transcendent state of affairs as exemplifying this property by the external object. 3.2 Existence as a “real predicate” I am aware that the construal of existence as a kind of property may strike you as contradicting some very clear statements from Brentano’s and Marty’s works. This reaction is partially justified. True enough, in his official writings Brentano always protested against interpreting the existential judgement as a predication of existence. (Cf. e.g. Brentano 1874/2008, 231) The existential acceptances and rejections are not predications, but rather mental attitudes toward a presented object, so that, surprisingly enough, in Brentanian existential judgements we don’t use the concept of existence (or non-existence) at all. Of course we can build these concepts if we want, but the mode of their construction is rather special. According to Brentano, they are obtained not by means of a standard abstraction singling out certain aspects of the objects of our mental acts, but rather by a kind of reflection (that Brentano called “immanent perception”) directed at the mental acts themselves. The mental acts from which the concepts of existence and non-existence are distilled are acceptances and rejections which are evident, and as such, as Brentano puts this, are “intrinsically characterised as right”. According to Brentano “to exist” thus means roughly “to be an object of a possible right acceptance” and “not to exist” means correlatively “to be an object of a possible right rejection”. (Brentano 1889/2010, 78) Marty accepts this analysis. (Marty 1884, 173f.) So the concepts of existence and non-existence do not refer to any intrinsic properties of things but rather to the ways these things can be (correctly) conceived of by conscious subjects. In this sense it is perfectly right that “existence” is not a “real predicate” – one of few points where Brentano agrees with Kant. (Cf. Brentano 1874/2009, 233) This is without doubt a philosophically deep and highly interesting interpretation of these puzzling concepts, but it seems that it only fits well with Brentano’s official, non15 propositional theory of judgement. It seems that a consequence of this construal of existence and non-existence is that there principally cannot be any “objective”, or “mind-independent” existences and non-existences in the world. If Brentano’s analysis is correct, then the concepts of existence and non-existence involve by definition a reference to the subjective operation of accepting and rejecting. Now Brentano’s unofficial doctrine of propositional contents explicitly introduces objective existences and non-existences in the world and thus literally turns this picture upside down. We can no longer claim that there is nothing in the objective, mind independent world that could be correlated with our predication of existence and non-existence. Evidently there is something over and above the involved objects which corresponds to them: namely the existences and non-existences of objects; and consequently there are also no reasons why existence shouldn’t be construed as a perfectly “real predicate”. So it seems that Brentano’s doctrine of propositional contents hardly leaves any room for an interpretation not ascribing to existence and non-existence a status of some – strange, special, esoteric, to be sure – but perfectly predicable properties. This is the reason why I assume this interpretation here. To be fair to Brentano, it must be repeated once more that this doctrine of propositional contents in an explicit form is to be found only in his non-published works. Brentano never decided to publish it and probably was aware that it involves many aspects that cannot be easily reconciled with his official non-propositional views. It was an irony of history that not only Marty, but in fact the majority of Brentano’s students tended to consider this theory of propositional contents as his official view.16 3.3 False-makers Well, can such a theory of existential states of affairs help us in explaining the question of truth-making? Of course within this ontological framework we can have a more “isomorphic” picture. For example for true negative judgements we have now non-existences of objects as 16 Incidentally, this interpretation of Brentanian existential states of affairs seems to support the Meinong-style criticism of his existential reduction. One of his arguments was that every composed nominal object has to involve something like an exemplification-nexus and is thus a state of affairs in disguise. Now we see that the only difference between Brentanian propositional contents and his nominal objects consists in the fact that the former ones exemplify strange properties of existence and non-existence, while the latter ones have only “standard” properties. But the fundamental ontological structure – that of exemplification of some properties – is in both cases exactly the same. So it seems that if Brentano’s composed objects are indeed nominal entities, then his existences and non-existences should be nominal as well; and if, on the contrary, his existential contents are really propositional, then also each composed (apparently nominal) object is in fact a propositional entity. 16 their truth-makers. Just as for every true acceptance of the form “+A” there is a state of affairs of the form that A exists (or the existence of A, as Brentano preferred to say), for every true rejection of the form “–A” we have a state of affairs of the form the non-existence of A. Now consider the case of a false acceptance. It goes without saying that if the judgement “+A” is false, then the judgement “–A” must be true. As we have seen, this judgement has now its truth-maker (namely the non-existence of A), and this truth-maker has a structure that in certain respects corresponds also to the false sentence “+A”. It involves the same nominal object and differs only in connecting it with non-existence instead of existence. This situation can be illustrated by the following scheme: E! encoding *E! non-existenceproperty ϕ identitfying property encoding exemplification immanent content Judgement + exemplification immanent object Presentation no REPR-relation and no intentional reference transcendent content reference object subject The illustrated judgement (that a certain particular Christmas tree, call it A, exists) is false, which means that it has no truth-maker. But within the framework of Marty’s theory it can nonetheless be correlated in a natural way with the state of affairs that A doesn’t exist. This state of affairs consists in exemplifying, by the Christmas tree in question, the property of non-existence, symbolised here as “*E!”. It makes the contradictory judgement “A doesn’t exist” true, and can be thus regarded as a false-maker of the judgement “A exists”. (The 17 immanent content of this judgement encodes the existence-property that has been symbolized as “E!”.) A false rejection “A doesn’t exist” has a quite similar intentional structure. It is false, so it has no truth-maker, but it has a false-maker: the existence of A which makes true the contradictory judgement “A exists”. Marty himself didn’t write about false-makers in this sense, but it is clear that we can find them in his rich ontology. In the following parts of this paper they will prove to be quite important for the formulation of Marty’s late theory. So in this respect Marty’s theory differs radically from Brentano’s official view outlined in section 2. According to Brentano’s non-propositional theory we need truth-makers (nominal objects) only for true acceptances. True rejections are true in virtue of correlative acceptances being false (i.e. in virtue of correlative acceptances’ not having truth-makers). Similarly there are no external correlates for false acceptances. We can find them only for false rejections. Within the framework of the ontology of immanent and transcendent propositional contents, we are therefore able to construct a much more “systematic” theory of truth- and false-making, but to be honest, the nature of this advantage seems to be mainly aesthetic. If we agree with Brentano and his followers, that all states of affairs have simple existential form, then this ontology of truth-making seems not to have more explanatory power then the elegant (Pseudo-)Brentanian theory from section 2.17 The situation changes of course, if we accept a more substantial, Meinongian criticism of the non-propositional theory of judgement. If this criticism is sound, then the majority of our truth-bearers simply cannot be made true by nominal objects, and states of affairs become very important. But, as already mentioned, this issue was irrelevant in Marty’s eyes. 17 There is a further aspect of this theory of states of affairs that I put aside in this paper. Brentano and Marty introduced also iterative states of affairs that should explain the semantics of judgements resulting from iterations of acceptances and rejections (like “The non-existence of Pegasus exists”, “The non-existence of the non-existence of a round square doesn’t exist” etc.) According to their theory, such judgements are made true by the correlative iterative contents (in our case by the existence of the non-existence of Pegasus and the nonexistence of the non-existence of the non-existence of a round square). Also modal judgements should be in a similar fashion made true by the appropriate iterative propositional contents. The truth-maker for the judgement “A round square doesn’t exist necessarily” would be for example the necessity of the non-existence of a round square. In Brentano 1889/1930 we find some exercises in the ontology of such iterative contents. While the semantics of non-modalised judgements is rather straightforward, so that it can be easily imagined how it works, the problem of modalities poses some difficulties that definitely go beyond the scope of this paper. For these reasons I restrict myself here to the non-modal and non-iterative judgements. 18 To close this section let me contrast the views of truth- and false-making of our (Pseudo-)Brentano and Marty with a stronger ontological position which, in addition to obtaining states of affairs, postulates also non-obtaining ones. The only entities figuring as truth- or false-makers in the Brentanian theory presented in section 2 were existing objects. From the history of philosophy we know, that it would be in principle possible to introduce a non-existing object for every non-veridical presentation. It was of course the position of Meinong, but Brentano never accepted it in his official writings.18 Also Marty uses in his theory of truth-making only obtaining propositional contents (he uses, like Brentano, the word “existing” rather than “obtaining” also in reference to propositional contents). So according to Marty, if a judgement is true, it always has a truthmaker, and if it is false, it has a false-maker, but there are no non-obtaining false-makers for true judgements and no non-obtaining truth-makers for false judgements. The position introducing such non-obtaining truth- and false-makers was held by the late Meinong. He introduces a mind-independent propositional correlate (called “Objective”) for every possible propositional attitude and divides the realm of these correlates into obtaining and non-obtaining ones. Under the assumption of this rich ontology all true judgements have obtaining truth-makers and non-obtaining false-makers, while all false judgements have obtaining false-makers and non-obtaining truth-makers. The following table summarises these analogies and differences:19 Judgements +A Brentano truthfalsemaker maker existing Marty truthFalsemaker maker obtaining true –A obtaining +A Obtaining false –A existing Obtaining Meinong truthfalsemaker maker obtaining nonobtaining obtaining nonobtaining nonobtaining obtaining nonobtaining obtaining 18 Although he toys with this idea in his unpublished Logic Lectures (EL 80). Cf. Chrudzimski 2001, 41f. 19 For the sake of comparison I assume here that all Meinongian judgements have Brentanian existential form. This is not true. In fact Meinong’s criticism of Brentano’s non-propositional theory of judgement was one of the most important steps leading him to his theory of objectives. For the details of Meinong’s consequently propositional theory of intentionality, see Chrudzimski 2007, chapters 4 and 5. 19 4. Marty’s late theory As is well known, after 1904 Brentano changed his ontological views and became a “reist”. Surprisingly for his students who still enjoyed the rich ontology of his lectures, he began to argue that there are no other entities than those belonging to the category of things.20 Marty agreed only partially. Between 1904 and 1908 he rejected all immanent entities (both nominal and propositional), but insisted that the transcendent states of affairs are indispensable, if we have to defend the objectivity of truth. The late Marty reduces thus his ontology of truthbearers, but not that of truth-makers. Marty’s arguments for the rejection of immanent entities are in principle the same as the arguments of the late Brentano. He said that the ontology of immanent entities doesn’t explain the phenomenon of intentional directedness, because in fact we are never intentionally directed at immanent entities. (Marty 1908, 395) Now immanent entities weren’t introduced without reasons. They were designed to play the role of contents (or Fregean senses) of our thoughts and to explain the puzzles of objectless presentations and non-analytic identities (like that between the Morning Star and the Evening Star). (Cf. Chrudzimski 2001, 29, 38) According to the doctrine of immanent entities what is missing in an “objectless presentation” is only its transcendent object. The immanent object is still here and so we can understand how an objectless presentation of a centaur can still be intentional. Further, the difference between different presentations having the same external object (like that of the Morning Star and of the Evening Star) consists precisely in involving different immanent objects. The immanent object of the first presentation encodes the property of being the Morning Star, while the immanent object of the second presentation encodes the property of being the Evening Star; and these are two quite different properties in spite of being exemplified by one and the same transcendent object. But now Marty has no immanent entities at his disposal. So what can be his explanation of these classical puzzles? His solution goes as follows: Presentation is a real process in a subject’s mind. When it takes place, the mind acquires a “relational characteristic” (relative Bestimmung) consisting in being capable to enter into a relation of the “ideal adequacy” with a certain object.21 20 Brentano’s things are concrete and individual, but not necessarily spatio-temporal. A spatially non-extended Cartesian ego or God as construed by most theists would be things in Brentano’s sense. 21 Cf. “Das Vorstellen ist ein realer Vorgang in der Seele, an welchem sich als nichtreale Folge knüpft, daß – falls dasjenige, was man das darin Vorgestellte nennt, existiert – die Vorstellende Seele zu ihm in eine eigentümliche Relation tritt, die sich etwa als eine ideelle Ähnlichkeit oder Adäquatheit mit demselben 20 So if I think of a centaur, I acquire a relational characteristics which is such that if there were a centaur in the world, I would stand in the relation of the ideal adequacy to it. The object to which the subject would stand in this relation of the ideal adequacy (provided it existed) is precisely what we call the object of this presentation. According to the first theory this object exists always “in the subject’s mind” as an immanent entity and a presentation consists precisely in having it before one’s mind. According to the new theory this object need not to exist and presentation consists only in the possibility of standing to it in the relation of the ideal adequacy, if it existed. Marty’s formulations give us basically two possibilities of interpretation of his late theory: (i) an adverbial and (ii) a counter-factual one. 4.1 An adverbial interpretation The adverbial approach to intentionality, first expressed by C. J. Ducasse and propagated in some works of Roderick Chisholm, claims that instead of introducing strange entities, like sense data or intentional objects, in explaining the phenomenon of intentionality we should rather take seriously the idea that an intentional mental state is at bottom a certain property of mind and the differences in the objectual reference must be therefore realized by certain modifications of this property. Now such modifications of properties (second-order properties) are normally expressed by means of adverbs. If I run, I always run somehow (quickly, slowly etc…), and similarly if I present (i.e. have a mental state of presentation), I always present somehow. In the traditional philosophical discourse we tended to say that I always present something, which led us inter alia to the (in)famous theory of immanent objects. But according to the partisans of the adverbial approach, we should take the form “somehow” seriously. What specifies the objectual reference of our presentations is not their relation to some strange entities but rather their second-order properties expressible by means of adverbs. To think of a centaur is not to have an intentional centaur before one’s mind; it is to think centaurly, to imagine something red is to imagine redly, to dislike George W. Bush, it to dislike George-W.-Bushly, and so on.22 bezeichnen läßt. [...] [M]it Rücksicht darauf mag man das Vorstellen geradezu einen Vorgang möglicher oder wirklicher ideeller Verähnlichung oder mentaler Adäquation mit einem Objekte nennen.” (Marty 1908, 406) 22 Cf. “The hypothesis [...] is that ‘blue’, ‘bitter’, ‘sweet’, etc., are names not of objects of experience, nor of species of objects of experience, but of species of experience itself. What it means is perhaps made clearest by saying that to sense blue is then to sense bluely, just as to dance waltz is to dance ‘waltzily’ (i.e., in the manner called ‘to waltz’) [...].” (Ducasse 1951, 259) 21 Now it seems that Marty’s second theory can be easily interpreted along the lines of the adverbial approach. A mental state (say a presentation) is according to Marty a “relational characteristic” of a mind. Provided the appropriate object exists in the world, this characteristic generates (together with the properties of the object in question) the relation of the “ideal adequacy” between the mind and the object. This means that the relation of the ideal adequacy is an internal relation supervenient on the monadic characteristics of the subject and the state of affairs. Marty calls such relations “grounded”, and says explicitly that the relation of the ideal adequacy between the mind and the corresponding external reference object belongs to the group of such grounded relations. (Marty 1908, 408) Here is the adverbial picture of a presentation: ϕ’ exemplification ϕ adverbial modification INTENTIONAL CORRELATION identitfying property exemplification Ω mental property of having a presentation exemplification IDEAL ADEQUACY intentional reference subject reference object The conscious subject on the left has a presentation of a Christmas tree. According to the present approach, this amounts to having a peculiar double-level mental property. First of all the subject has a presentation, which means that he or she exemplifies the mental property of having a presentation (or presenting) symbolised as Ω. Second this presentation exemplifies a higher-order property of being “of a Christmas tree”. This is the adverbial modification labelled as ϕ’. It is assumed that an adverbial modification of this kind specifies the 22 correlative identifying property (symbolised here as ϕ). There must be thus a certain definite relation between these two properties, that I called the relation of intentional correlation.23 How does it work in the case of a judgement? Here is the picture: E!’ E! ϕ’ ϕ existenceproperty exemplification exemplification adverbial modifications Ω INTENTIONAL CORRELATION identitfying property exemplification presentation exemplification IDEAL ADEQUACY intentional reference subject transcendent content reference object The only difference on the subjective side is that the mental property of having a presentation has now two adverbial modifications: the usual ϕ’, specifying the identifying property ϕ, and a little strange E!’ (intentionally-correlative to E!). The latter is meant to secure that in the judgement under discussion the presented ϕ-object “is supposed to exist” (i.e. has been accepted in Brentano’s sense). Consequently the intended correlate on the objective side is no longer a simple ϕ-object, but the propositional existence of a ϕ-object. Compared with Marty’s theory of immanent entities from the previous section the adverbial theory proposes without doubt a less extravagant ontology. We have here only 23 In Cesalli 2008 it is argued that Marty’s relative Bestimmung should be interpreted not as a monadic property of a conscious subject, but rather as a particular type of relation. Cesalli sometimes calls it a possible relation, which introduces still another problem – that of possibilia. As should be clear from this section I don’t think that this is a correct reading of the late Marty (section 4.3 proposes only a certain linguistic ersatz for it). However the late theory of Brentano can be interpreted as introducing strange non-extensional relations of a similar kind. Cf. Chrudzimski 2001, 239. 23 standard properties and standard exemplifications. Consequently in the last picture we find no “clouds” with strange, mind-dependent entities inside. This is the reason why so many philosophers find the adverbial theory of intentionality attractive. 4.2 The importance of false-makers But the adverbial theory of intentionality has also its dark side. Its weakest point is the relation between the relational characteristic, marked in our picture as ϕ’, and the identifying property ϕ – the relation of intentional correlation. The nature of this relation is absolutely crucial to the understanding of how the adverbial mechanism works. Recall, the general picture is this: To refer intentionally to a ϕ-object consists in exemplifying the relational property ϕ’, and that’s all that must be done. In particular there need not to be any ϕ-object in the world. I symbolised the relational property securing intentional reference to a ϕ-object as ϕ’, which suggests that the relation between the identifying property of a (potential) target object and the relational property of the conscious subject must be something very straightforward and directly understandable. In doing this I followed the policy of the partisans of the adverbial theory, whose use of the simple adverbial suffix “-ly” smuggles in the same suggestion. Remember: to perceive something red is to perceive redly, to believe that snow is white – to believe white-snowly etc. The fact is however, that we have neither any precise idea of the nature of those adverbial modifications, nor the slightest hints concerning the character of the relation of intentional correlation. All we know about the property of perceiving redly is that it is a mental property that secures an intentional directedness at a red object, and all we know about the relation of intentional correlation is that it is the relation obtaining between any of the possible mental adverbial modifications and a correlative identifying property (i.e. for example between properties expressed by words “redly” and “red”). It seems that both adverbial modifications and the relation of intentional correlation are perfect examples of what is called “theoretical” or “postulated” entities. They are not our “data”, they never appear before our eyes (or mind), but are introduced as parts of a hidden mechanism that is designed to explain some of such “data”. Now it seems that this fact has really bad consequences for the adverbial theory in general. To see the problem, consider what happens, when the intentional state doesn’t hit any target in the world. The main idea of the adverbial approach is that we need in this case no “acting person” who would be able to play the role of the reference entity. In particular we 24 have no Brentano-style immanent objects. What we have are only some double-level properties of the subject. But we have seen above that the conscious subject in question has no epistemic access, either to the property ϕ’, or to the relation of intentional correlation; and it seems that we have to require both, if the adverbial theory has to work at all. To understand this we must only realise, that to be able “to know” what is his or her proper reference object, the subject in question has to “reconstruct” the identifying property ϕ from the property ϕ’ and the relation of intentional correlation. The only “information” the subject has concerning the identifying property in namely the following: “the property that would stand in the relation of intentional correlation to the property that is currently the adverbial modification of my presentation”.24 Well, this is a general difficulty of the adverbial approach in its typical form, and I think that it is indeed fatal, but it turns out that it doesn’t affect Marty’s late theory, because of his rich ontology of propositional contents. To repeat, the problem of the adverbial theory is that in the case of a mental state that misses its target the identifying property ϕ has to be reconstructed from the adverbial modification and the relation of intentional correlation. It has to be reconstructed, because there is no property ϕ in the world around us. If we had the identifying property ϕ somewhere, there would be no need of reconstructing it. This means, first of all, that no Platonist will ever have any problem with this difficulty. For him or her, the property ϕ exists, even if it is not exemplified by any individual. Second, also for a Meinongian, this difficulty is absolutely harmless. A Meinongian can provide you with as many ϕ-objects as you want. Sometimes they will be all non-existent, but it’s no problem in the present context. Now Marty was neither a Platonist nor a Meinongian – he believed neither in Platonic non-exemplified properties nor in Meinongian non-existent objects “beyond being and non being” –, still also in his universe we can always find the required identifying property ϕ. The reason is that beside obtaining existences Marty has also obtaining non-existences. Provided a ϕ-object exists, we have no problem with the property ϕ. It is exemplified by this object. But even if there is no ϕ-object in the world, there is still an obtaining non-existence of a ϕ-object, and so there is also the property ϕ as a constituent of this non-existence. 24 One can think that these epistemic factors have no relevance here, but in fact they have. If we start to analyse the classical Fregean problems that triggered the development in theory of intentionality (particularly the problem of non-analytic identities), it turns out that some kind of privileged epistemic access to the mechanism of intentional reference must be secured if the theory has to work at all. I have no place here to go into details. For a more developed argument see Chrudzimski 2005b. 25 Let me illustrate this by a schematic picture of a false positive judgement according to the adverbial construal of Marty’s late theory: E!’ *E !ϕ ϕ’ Non-existenceproperty exemplification exemplification adverbial modifications Ω INTENTIONAL CORRELATION identitfying property exemplification presentation exemplification no ideal adequacy no intentional reference subject transcendent content reference object The depicted judgement is false, which means that there is no intentional reference and no relation of the ideal adequacy between the subject’s mind and the reality. Nonetheless there is a certain state of affairs that can be in a natural way correlated with the judgement – namely its false-maker. This state of affairs is of course the non-existence of a ϕ-object. Strangely enough, it “involves” a ϕ-object, and thus also the property ϕ, as its “part”. Consequently there is still room for the relation of intentional correlation. It holds, exactly as in the case of a true judgement, between ϕ’ and ϕ. Under assumption of Marty’s rich ontology we can thus always find the required identifying property; and the epistemic inaccessibility of the adverbial modification as well as the relation of intentional correlation need not to be interpreted as a fatal objection to his theory in its adverbial reading. The adverbial modification and the relation of intentional correlation can be regarded as parts of a “hidden mechanism” of intentional reference that leads us eventually to the identifying property, which alone stands before the subject’s mind. And concerning the second adverbial modification (E!’) there are of course many existences 26 in the world intentionally correlated with the adverbial modification E!’. I didn’t depict this relation in my scheme only for the sake of transparency. It goes without saying that in the case of a false negative judgement the situation is absolutely analogous. Also in this case we have a false-maker – the existence of the object in question. Also in this case the involved object exemplifies the required identifying property, and also in this case there are of course many non-existences in the world with which the adverbial modification *E!’ is intentionally correlated. The existence of false-makers proves thus to be a quite important feature of Marty’s late theory in its adverbial reading. 4.3 A counter-factual interpretation To close I want to signalise another possible interpretation of Marty’s late theory, without developing it in detail. In his explanations of what it means for an “objectless” intentional state to be intentional, the late Marty often employs counter-factual discourse. According to this explanation, to say that an intentional state “has a certain object” (say A) is nothing other than to say that it would stand in the relation of the ideal adequacy to A, if A existed.25 This kind of explanation gives us a possibility of another interpretation of his late position. We have seen above that the main difficulty of the adverbial interpretation consists in the fact that both the adverbial modification ϕ’ and the mediating relation of intentional correlation are “postulated” entities, hidden from the subject’s eyes. This typically has the unpleasant consequence that in the case of a non-veridical intentional state there seems to be literally nothing that can stand before the subject’s mind. The subject “sees” neither the property ϕ’, nor the relation of intentional correlation; and the property ϕ itself cannot be seen in the first place, because it doesn’t exist. Now the identifying property ϕ is precisely what the subject should be aware of. When I imagine a centaur, it is the property of being a centaur that I have before ma mind. Any theory of intentionality that denies this fact deserves the name of a non-theory par excellence. We have also seen that Marty’s rich ontology can save him from the label of a non-theorist of intentionality. As he has in his ontological universe both existences and non-existences, he has all thinkable properties at his disposal. 25 Cf. “Wenn ich eine Farbe vorstelle, so ist die Farbe das Objekt dieses Vorstellens, d.h. wenn eine Farbe wirklich existiert, so steht mein Vorstellen zu ihr in einem eigentümlichen Verhältnis ideeller Adäquation [...]. Aber auch wenn meiner Farbenvorstellung nichts in Wirklichkeit adäquat ist, sage ich doch, sie habe ein Objekt, weil, wenn eine wirkliche Farbe bestände, sie notwendig in diesem Verhältnis des Entsprechens zu jener Vorstellung stehen müßte.” (Marty 1916, 44) 27 Now the reference to the counter-factual discourse opens another way. Assume for a moment that we operate within the framework of the Brentanian ontological universe from section 2, containing only existing objects and no states of affairs. Assume further that we reject all immanent entities. In this universe there is of course no place where the property of being a centaur can be found. But now imagine that we use the counter-factual discourse in defining the adverbial property of presenting centaurly: S presents centaurly if and only if S would stand in the relation of the ideal adequacy to anything that were a centaur. Notice that this definition involves the reference to a centaur (and thus to the identifying property of being a centaur), and thus we can no longer claim that having relational characteristics in Marty’s sense, do not put identifying properties before the subject’s mind. Now they involve identifying properties in their very definition. But the big question is of course, how this whole mechanism is supposed to work. We already agreed that in our entire ontological universe no property of being a centaur can be found. So how can it figure in any definition? It seems that the only way to have this property “before one’s mind” is indeed to “reconstruct” it from the property of presenting centaurly and the relation of intentional correlation, which puts us back into our old predicament. Now the solution of this dilemma is quite simple but not uncontroversial. To cut the Gordian knot, one has only to assume that the counter-factual discourse is irreducible in defining the relational properties in Marty’s sense. The presence of the counter-factual idiom must be accepted as a primitive feature of our philosophical grammar. Sadly enough, says a partisan of this interpretation, some philosophically important things cannot be expressed in a language deprived of counter-factual operators. Well, this was the simple side of the counter-factual interpretation. The controversial one emerges as soon as we try to give an ontological interpretation of this claim. It seems that the only reasonable ontological articulation of the irreducible character of the counter-factual discourse is the assumption to the effect that the underlying ontology is irreducibly dispositional. In application to our problem it would mean that Marty’s relational characteristics are to be construed as ontologically primitive dispositions, dispositions that cannot be reduced to any combination of purely categorical properties. To accept this explanation we must be thus prepared to swallow such primitive dispositions in the world, and so it will be unacceptable for all those who believe that all 28 dispositions must have their categorical basis. Concerning the historical Marty, it seems that he wasn’t fully aware of the possible ontological commitments resulting from the use of counter-factuals. In particular he never addressed the question whether there are irreducibly dispositional layers of reality. We have a quite similar situation in the late writings of Brentano. In getting rid of all entities not belonging to the category of things he often used counter-factual constructions without any serious attempt to analyse their semantics. Apparently he believed that the counter-factual discourse is ontologically innocent. References Alston, William P. (1995) A Realist Conception of Truth, Ithaca and London: Cornell University Press. Armstrong, David Malet (2004) Truth and Truthmakers, Cambridge: Cambridge University Press. Brentano, Franz (1874/2008) Psychologie vom empirischen Standpunkte, in: F. Brentano, Psychologie vom empirischen Standpunkte. Von der Klassifikation der psychischen Phänomene (Sämtliche veröffentlichte Schriften, hrsg. von T. Binder, A. Chrudzimski, Bd. I), Frankfurt a.M., Ontos-Verlag, 1-289. Brentano, Franz (1889/2010) Vom Ursprung sittlicher Erkenntnis, in: F. Brentano, Schriften zur Ethik und Ästhetik (Sämtliche veröffentlichte Schriften, hrsg. von T. Binder, A. Chrudzimski, Bd. III), Frankfurt a.M., Ontos-Verlag, 19-98. Brentano, Franz (1889/1930) “Über den Begriff der Wahrheit”, in: Brentano 1930, p. 3–29. Brentano, Franz (1891/1982) Deskriptive Psychologie, hrsg. von Roderick M. Chisholm, Wilhelm Baumgartner, Hamburg: Meiner. Brentano, Franz (1930) Wahrheit und Evidenz, hrsg. von O. Kraus, Hamburg: Meiner. Carnap, Rudolf (1934) Logische Syntax der Sprache, Wien: Springer. Cesalli, Laurent (2008) “Relative Bestimmung. Une relation martyienne”, in: C. Erismann, A. Schniewind (éds.), Compléments de substance. Études sur les propriétés accidentelles offertes à Alain de Libera, Paris: Vrin 2008, 216-229. Chrudzimski, Arkadiusz (2001) Intentionalitätstheorie beim frühen Brentano, Dordrecht / Boston / London, Kluwer Academic Publishers. Chrudzimski, Arkadiusz (2004) Die Ontologie Franz Brentanos, Dordrecht / Boston / London: Kluwer Academic Publishers. 29 Chrudzimski, Arkadiusz (2005a) Intentionalität, Zeitbewusstsein und Intersubjektivität. Studien zur Phänomenologie von Brentano bis Ingarden, Frankfurt a. M.: OntosVerlag. Chrudzimski, Arkadiusz (2005b) “Brentano, Husserl und Ingarden über die intentionalen Gegenstände”, in: A. Chrudzimski (ed.), Existence, Culture, and Persons: The Ontology of Roman Ingarden, Frankfurt a. M.: Ontos-Verlag, 83–114. Chrudzimski, Arkadiusz (2007) Gegenstandstheorie und Theorie der Intentionalität bei Alexius Meinong, Dordrecht / Boston / London: Springer. Chrudzimski, Arkadiusz (2009) “Sachverhalte, Objekte und Supervenienz. Brentano, Marty und Meinong”, Brentano Studien, 12 (2006/9), 99–119. Ducasse, C. J. (1951) Nature, Mind, and Death, LaSalle, Ill.: Open Court. Dummett, Michael (1976) “What is a Theory of Meaning? II”, in: G. Evans & J. McDowell (Eds.), Truth and Meaning, London: Oxford University Press. Johansson, Ingvar (1989) Ontological Investigations, London: Routledge. Marty, Anton (1884) “Über subjectlose Sätze und das Verhältnis der Grammatik zu Logik und Psychologie” I–III, Vierteljahresschrift für wissenschaftliche Philosophie, 8 (1884), 56–94, 161–92, 292–340; reprinted in Marty 1916–1920, Bd. II/1, 1–35, 36– 62, 62–101. Marty, Anton (1908) Untersuchungen zur Grundlegung der allgemeinen Grammatik und Sprachphilosophie, Halle: Niemeyer. Marty, Anton (1916) Raum und Zeit, aus dem Nachlasse des Verfassers hrsg. von J. Eisenmeier, A. Kastil und O. Kraus, Halle: Niemeyer. Marty, Anton (1916–1920) Gesammelte Schriften, hrsg. von J. Eisenmeier, A. Kastil and O. Kraus, Bd. I/II, Halle: Niemeyer. Meinong, Alexius (1892) „Rezension von: Franz Hillebrand, Die neuen Theorien der kategorischen Schlüsse“, in: Gesamtausgabe, vol. VII, 197–222. Meinong, Alexius (1902) Über Annahmen, 1st ed., Leipzig: Johann Ambrosius Barth [partially reprinted in: Gesamtausgabe, vol. IV, 385–489]. Meinong, Alexius (1969–78) Gesamtausgabe, eds. R. Haller et al., Graz: Akademische Druck- und Verlagsanstalt. Mulligan, Kevin, Simons, Peter and Smith, Barry (1984) “Truth-makers”, Philosophy and Phenomenological Research 44 (1984), 287–321. 30 Peirce, Charles Sanders, (1868) “Some Consequences of Four Incapacities”, in: Peirce 1992, 28–55. Peirce, Charles Sanders (1878) “How to Make our Ideas clear”, in: Peirce 1992, 124–141. Peirce, Charles Sanders (1992) The Essential Peirce, Selected Philosophical Writings, Vol. 1 (1867–1893), Bloomington and Indianapolis: Indiana University Press. Popper, Karl (1935) Logik der Forschung, Wien: Springer. Putnam, Hillary (1981) Reason, Truth, and History, Cambridge: Cambridge University Press. Putnam, Hillary (1995) Pragmatism. An Open Question, Oxford: Blackwell. Ramsey, Frank P. (1927/1990) “Facts and Propositions”, in: F. P. Ramsey, Philosophical Papers, ed. by D. H. Mellor, Cambridge: Cambridge University Press 1990, 34–51. Russell, Bertrand (1910) “Some Explanations in Reply to Mr Bradley”, Mind N.S, 19, 373– 378. Sellars, Wilfrid (1967) Science and Metaphysics: Variations on Kantian Themes (The John Locke Lectures for 1965–66), London: Routledge and Kegan Paul. Srzednicki, Jan (1965) Franz Brentano’s Analysis of Truth, The Hague: Nijhoff. Tarski, Alfred (1933) Pojęcie prawdy w językach nauk dedukcyjnych [The Concept of Truth in the Formalized Languages]; English version in Tarski 1956, 152–277. Tarski, Alfred (1956) Logic, Semantics, Metamathematics. Papers from 1923 to 1938, Oxford: Clarendon Press. Zalta, Edward N. (1988) Intensional Logic and The Metaphysics of Intentionality, Cambridge, Mass.: The MIT Press. 31