From the ALWS archives: A selection of papers from the International Wittgenstein Symposia in Kirchberg am Wechsel

1. The phenomenal concept strategy

One of the most famous objections to Jackson’s knowledge argument (Jackson 1986) is the so-called two modes of presentation-reply. The basic idea of this reply – which is the possibility that one single, ontological fact can be known under different modes of presentations – can be easily formulated on the level of concepts. This move leads to the notion of phenomenal concepts on the one hand and the notion of physical concepts (understood in the widest sense) on the other hand. These two sorts of concepts then are treated in analogy to standard cases of co-reference. Hence, according to the two modes of presentation-reply the brilliant scientist Mary possessed all physical concepts, when being confined to her achromatic room, but gained new phenomenal ones, when enjoying her first colour-experience. Obviously, only type-B-materialist (Chalmers 1997), which grant that phenomenal concepts can not be a priori deduced from physical concepts, can adopt the physicalistic phenomenal concept strategy (Stoljar 2005). In other words: physicalists, who intend to save an ontological materialism by granting just a conceptual or epistemic gap, developed this interpretation of the knowledge argument to reach their target.

The physicalist phenomenal concept strategy is based on the idea that the particularities of phenomenal concepts can explain why one can not deduce them a priori from physical concepts, although both sorts of concepts pick out one and the same ontological (ex hypothesi physical) referent. Hence, with regard to Mary it can be said that no metaphysical entities such as qualia have to be invoked to explain the scientist’s new knowledge – it suffices to point out the uniqueness of phenomenal concepts. For this strategy to work, the decisive features of phenomenal concepts have to be elaborated. These particularities will have to explain why phenomenal concepts are conceptually isolated (Carruthers, Veillet 2007) from other concepts, but still pick out physical referents.

In the following I will demonstrate that if we take the uniqueness of phenomenal concepts seriously, we have to conclude that they refer to phenomenal entities and therefore the physicalist phenomenal concept strategy fails. I will start working out the crucial particularities of phenomenal concepts: one particularity concerns the concept-acquisition and the other the very nature of such concepts. Importantly, both particularities of phenomenal concepts are such that they indicate phenomenal referents. In a second step, I will analyze one interpretation of phenomenal concepts which seems to describe the crucial particularities of these concepts adequately: Papineau’s quotational account of phenomenal concepts (Papineau 2002, 2007). A detailed examination of this account will reveal two possible interpretations: the first interpretation is similar to the herein presented account and therefore leads to a dualistic conclusion. The second interpretation fails to explain the decisive features of phenomenal concepts; such as their semantic stability and the closely linked fact of carrying information about qualitative experiences. Hence, Papineau has to choose between accepting that phenomenal concepts do refer to phenomenal referents or defending a view of phenomenal concepts which leave the crucial particularities of phenomenal concepts and therefore also the Mary-scenario unexplained.

2. The encapsulation relation explains the particularities of phenomenal concepts

Let me start my investigation analyzing the particularities of phenomenal concepts. Regarding the concept-acquisition, the knowledge argument famously illustrated that we can gain phenomenal concepts only under the condition of attentively experiencing their referents. In other words: one has to stand in the extraordinary intimate relationship of acquaintance with the referent a phenomenal concept picks out. Hence, when Mary leaves her achromatic environment, sees for the first time the blue sky and is attentively aware of this colour-experience, she gains a new phenomenal concept. Let me explain this process in more detail: the brilliant scientist, who is aware of her very first blue-experience, discriminates this experience from all other current experiences. In my opinion it is this act of attentive discrimination which immediately yields a concept referring to this particular, isolated experience. The close link between an experience and the gained conception of it is a crucial point for my further argumentation.

Regarding the nature of phenomenal concepts, a careful analysis reveals an encapsulation relation between these concepts and the referents they pick out. The notion of an encapsulation relation can be considered as fundamental for the presented account. It is based on the idea that the experience itself is the core of the phenomenal concept referring to it. This fact can be explained by the special way of gaining these concepts: when Mary discriminates a new experience she is acquainted with, this process of isolation implies giving the experience itself a conceptual structure and hence forming a phenomenal concept which encapsulates the very experience itself. Obviously according to this account, both the concept and the referent are occurrences in the subject’s mind. The intimate link of encapsulation of the referent in the concept has very particular roots and consequences:

One crucial root of the encapsulation is the self-presenting character of the referent, which enables the direct reference of the concept. It is precisely the fact that an experience is self-presenting, i.e. that it serves as its own presentation, which is responsible for our acquaintance and discriminative awareness of it and hence points towards the close link between experience and phenomenal concept.

The decisive consequences of this account are the following: phenomenal concepts pick out their referents directly and in all possible worlds – facts which are due to the internal constitution of encapsulation. Importantly, since the reference of phenomenal concepts is fixed by their constitution and not by external factors, they carry essential information about their referents. Taking the Mary-scenario into account it becomes evident that the relevant information has to be about the qualitative character of experiences because it is precisely this sort of information the scientist lacked in her achromatic room and gained when looking at the sky.

3. Examples of alternative accounts of phenomenal concepts

In my opinion solely the encapsulation relation can explain the particularities of phenomenal concepts. Consider, for example, the fact that released Mary gains a new concept which importantly carries information about the very experience she is undergoing. No demonstrative account of phenomenal concepts, such as, for example, the one developed by Levin (Levin 2007), can capture this function of phenomenal concepts. Demonstrative concepts typically refer to the item currently demonstrated at and hence their referents differ from one use to another. Contrary to this, my account of phenomenal concepts makes them pick out their referent necessarily and in all possible worlds. Remember, a phenomenal red-concept should necessarily carry information about phenomenal redness to explain the Mary scenario and demonstrative concepts do not meet this constraint.

If we consider direct recognitional phenomenal concepts of the sort invoked by Loar (Loar 1997), we are confronted with another sort of problem: obviously our capacities to discriminate experiences outrun our capacities to recognize experiences. Suppose, Mary has an experience of the shade red21 parallel to shade red23 and can discriminate these two shades introspectively. Nevertheless, she may not be able to recognize these shades when she encounters them. According to the recognitional account of phenomenal concepts Mary has no phenomenal concept of red21 or red23, although she is attentively experiencing these shades and at this moment knows, what it is like to see them. I take this to be a quite implausible conclusion.¹

These considerations illustrate that no account of phenomenal concepts which neglects the intimate link between these concepts and their referents can successfully explain the particularities of the concepts Mary acquires because of her first colour experience. In addition, accounts which take phenomenal concepts and experiences as separate entities, related to each other only causally, face a further problem: as Balog (Balog, forthcoming) points out, on such accounts it is conceivable that a first-person’s application of a phenomenal concept is performed even in the absence of the experience it refers to – and this is quite an absurd way of treating phenomenal concepts. For this reasons, let me return to my thesis of phenomenal concepts encapsulating their referents.

4. The dilemma of Papineaus´s quotational account of phenomenal concepts

In the following I want to focus my attention on a physicalist account, which seems to share the herein elaborated interpretation, but draws physicalistic conclusions from this: the so-called quotational account of phenomenal concepts. Papineau developed this account in his book Thinking about consciousness (Papineau 2002), but changed some details in a recent article (Papineau 2007).

The quotational account is based on the assumption that phenomenal concepts embed experiences just as quotation marks embed words. If his analogy is worked out in detail, we will see why Papineau faces a dilemma: if his account is understood as a sort of real encapsulation, then he has to conclude that phenomenal concepts pick out phenomenal referents. The reasons for this conclusion are the following: if phenomenal concepts are interpreted as encapsulating their referents, then this unique reference relation has to be explained. According to my analysis, solely an explanation referring to the self-presenting character of phenomenal properties and our special acquaintance relation to them can do this explanatory work. If one wants to avoid this dualistic conclusion, she has to give a physicalistic account of how a concept can encapsulate and directly refer to a physical item and it seems mysterious how this can be done without invoking self-presenting (phenomenal) properties.

The remaining option is to interpret the quotational account as phenomenal concepts just using experiences without granting that they are a logical part of the concept itself. In fact, Papineau in his article “Phenomenal and perceptual concepts” (2007) doesn’t seem to believe anymore that the particularities of phenomenal concepts lie in a unique reference relation, but rather holds that they can be explained by the special (neuronal) vehicle in virtue of which the concept is realized. This suggests that the presence of the experience in the concept should be explained by a physical (neuronal) presence:

We can helpfully think of perceptual concepts as involving stored sensory templates. These templates will be set up on initial encounters with the relevant referents. They will then be reactivated on later perceptual encounters. (Papineau 2007, 114)

Obviously the “stored sensory template” has to be understood as a physical item. At this point some pressing questions arise: firstly, what is meant by “involving” these templates? If this phrase only points at simultaneous occurrence of concept and experience, then the concept doesn’t carry any information about the qualitative character of the experience. If the citation has to be understood as a constitutional relation, one may wonder a) how a physical item (as a neuronal template) can be encapsulated in the concept and b), how it can carry the relevant information. Ad a) it can be pointed out that on a physicalist account no primitive acquaintance relation can be invoked to explain this constitution and that neural templates are not introspectively accessible. Next, b) has to be explained in more detail: the information a phenomenal concept has to carry surely is not information about a neural state – otherwise Mary would have possessed this concept in her achromatic environment. A phenomenal concept has to carry information about the qualitative character of the experience and it is unclear how a physically understood template can do this, without recurring to phenomenal properties. A purely physical description of a (neuronal) template would obviously leave out precisely the sort of information a phenomenal concept has to carry to explain Mary’s situation. Therefore, if Papineau´s account of phenomenal concepts is interpreted as solely co-occurring with experiences or as involving physical items, then the decisive particularities of the concepts will not be explained adequately anymore.

5. Conclusion

I want to summarize my line of thought: in accordance with most phenomenal conceptualists I showed that the concepts involved in the Mary-case differ in several respects significantly from any other concept the scientist had before her release. But the central point of my analysis – which stands in contrast to target of the physicalist phenomenal conceptualist – was to argue that these differences are such that the new concepts refer (because of their internal structure) necessarily to phenomenal entities.

In a next step, I compared the elaborated account of phenomenal concepts with some physicalistic ones. I demonstrated that the basic assumptions of most physicalist phenomenal conceptualist (as Levin or Loar) can not explain the crucial particularities of phenomenal concepts. Then I focused the attention on the quotational account advocated by Papineau which at first glance seemed to describe these particularities adequately. But a careful analysis illustrated that also Papineau’s account has consequences which stand in contrast to the target the physicalist intends to reach: if it is understood as just involving physical items, then it can not meet the constraint of explaining the decisive particularities of phenomenal concepts; such as carrying information about the qualitative character of experiences. But if it is interpreted in accordance with the herein advocated encapsulation relation, then it has exactly the dualistic consequences the physicalist phenomenal conceptualist wants to avoid.