SORITES, ISSN 1135-1349

Issue #11. December 1999. Pp. 66-81.

Are There Mental Entities? Some Lessons from Hans Reichenbach

Copyright © by SORITES and Jeanne Peijnenburg

Are There Mental Entities? Some Lessons from Hans Reichenbach

by Jeanne Peijnenburg

0. Introduction

`The mental and the physical are not made for each other', wrote Davidson, echoing Brentano's famous thesis that the intentional idiom is irreducible. But if mental terms cannot be translated into physical terms, how can they be translated? What is the meaning of mental terms such as `belief', `desire', `intention'? Or to put it ontologically: what sort of entities are beliefs, desires, intentions?

These questions are core issues in the contemporary philosophy of mind, and the answers are many. Beliefs and desires have been related to actions, to brain processes and to computer programs. The relations in question have been pictured as reductions, as superveniences, as emergence relations, as type- and as token-identities. Nowadays there exists a whole gamut of different positions on the mental: there is physicalism, eliminativism, behaviourism, functionalism, parallelism, epiphenomenalism, interactionism, anomalous monism, and, last but not least, transcendentalism.

In this paper I propose to discuss Hans Reichenbach's views on the mental, especially his views on abstracta and illata, and compare them with some ideas of Carnap. It was Daniel Dennett who, while explaining his own views on the nature of mental entities, drew attention to Reichenbach's abstracta and illata in Dennett 1987 (cf. Dennett 1991a and Dennett 1991b). However, Dennett is not a historian of philosophy. He never aimed to present Reichenbach's theory in full detail, but only cited it in passing. As a result, the great potential of Reichenbach's ideas concerning mental entities still remains largely unnoticed.

Reichenbach is often considered to be an adamant logical positivist, propagating ideas on the mental that are far too `physicalistic'. Like Carnap, he is taken for an unrefined behaviorist who perceives the psychological realm as consisting only of gross stimuli and raw responses. In fact, however, his ideas on mental entities are quite sophisticated and by no means the crude positivistic products that some make of them. For instance, as we will see, Reichenbach makes ample room for private experiences and for the first person view, matters that after all are Fremdkörper in behavioristic theories of the more simple sort.

Hence I am going to undertake a journey that is rare in analytic philosophy: I propose to go back in time and examine ideas that are more than fifty years old. Analytic philosophers are in general not historically oriented. They tend to forget that an occasional excursion into the past may be worthwhile, especially when the jaunt includes a visit to one's very own roots, as is the case here. Carnap and Reichenbach are early representatives of analytic philosophy, and they commented thoroughly on the nature of abstract entities and the meaning of abstract terms. As we will see, a comparison of their views yields lessons about the mental that might still be of worth today.

1. Reichenbach: reduction and projection

In Experience and Prediction Hans Reichenbach distinguished between direct and indirect propositions. At first sight, the distinction is an unalloyed neo-positivistic product. Direct propositions are the familiar observation sentences capable of direct verification; indirect propositions are indirectly verified, which means that they are reducible to other propositions capable of direct verification.

The interesting question, of course, is what exactly `reduction' means here. What does it mean for example to say that a proposition about an event horizon, i.e. the border of a black hole where the escape velocity equals the speed of light, `is reducible to' a class of observation sentences? Cosmologists in the entourage of Stephen Hawking detect an event horizon by measuring electromagnetic radiation emitted from a shrinking star, and by comparing the measured signals with the predictions of quantum field theory and general relativity. The cosmologist's claim that in an event horizon the photons `hover', i.e. neither escape from the hole nor fall back into it, is based on various sentences concerning outcomes of measurements made with miscellaneous instruments. Each of those instruments, we assume, is placed on our planet, thousands of millions of miles removed from what they are observing: the happenings in an event horizon. What is the relation between the (indirect) statement that the photons in an event horizon hover, and the (direct) statements about results of measurements?

Reichenbach's answer to this kind of question is often put on a par with that of the early positivists. However, the differences are considerable. The early positivists regard any relation between direct and indirect statements as an equivalence: an indirect statement (IS) is true if and only if the set SD of direct statements is true, where SD can contain conjunctions, disjunctions, negations etc. Reichenbach, on the other hand, finds this view too simple. He points out that often IS has a surplus meaning compared to the meaning of the propositional function of the statements in SD. In those cases IS can be true while one or more statements in SD may be false, and vice versa. Hence the relation might be not an equivalence but a probability connection: IS probably implies SD and vice versa.

Reichenbach calls the probability connection a projection, and the equivalence relation a reduction. An example of a reduction is the relation between (1) «The species of wallabies has its home in Australia» and (2) «All wallabies descend from ancestors that lived in Australia» (the example is a modified version of Reichenbach's example). (1) is an indirect statement, for it contains indirectly verifiable terms: `the species' does not denote a concretum, and neither does `home'. It is however completely reducible to (2), which contains, besides logical terms such as `all', only terms that refer to what Reichenbach calls concreta, i.e. physical objects or processes supposedly accessible to direct observation. In Reichenbach's words, (1) denotes a non-concretum that is a reductive complex and the expressions in (2) refer to the internal elements of this complex (Reichenbach 1938, 110). Another example of a reduction is the relation between a wall and the bricks of which it is built. Every statement about the wall (the reductive complex) can be translated into a statement about the bricks (the internal elements). Of course, the bricks can only form a wall if they are arranged in a particular way: the wall is not dependent upon just bricks, but upon a certain configuration of the bricks. Thus Reichenbach says that the reductive complex is equivalent to the internal elements together with a «constitutive relation».

On the other hand, if an indirect statement is connected to direct statements through a projection rather than through a reduction, then the indirect statement denotes a projective complex and the direct statements refer to external elements. Reichenbach gives the following example of a projection:

We imagine a number of birds flying within a certain domain of space. The sun rays falling down from above project a shadow-figure of every bird on the soil, which characterizes the horizontal position of the bird. To characterize the vertical position also, let us imagine a second system of light rays running horizontally and projecting the birds on a vertical plane which may be represented by a screen of the kind employed in the cinemas. We have, then, a pair of shadows corresponding to every bird ... every proposition concerning the movement of the birds is co-ordinated with a proposition about the changes of the pairs of shadows. (Reichenbach 1938, 108).

In this example, every single bird is represented by a unique system of marks, in the sense that each movement of the bird corresponds to a movement of the shadows. The birds are however not identical to the shadow pairs, no matter how the latter are arranged with respect to each other. Instead, the birds are only projected on to the screen and the soil: they constitute projective complexes of which the shadows are the external elements. This means that no proposition about a bird is completely reducible to a proposition about a shadow pair, and hence that between propositions about the birds and propositions about the shadows only probability connections exist:

if we see the marks only, we may infer with a certain probability that they are produced by birds, and if we see the birds only, we may infer with a certain probability that they will produce the marks. ... there is no strict relation between the truth values of the co-ordinated propositions. The proposition about the birds may be true, and that about the marks may be false; conversely, the proposition about the birds may be false, and that about the marks may be true. (Reichenbach 1938, 109).

Projective complexes such as the birds are called illata, i.e. `inferred things' (Reichenbach 1938, 212) -- other examples of illata are radio waves, atoms, and all sorts of invisible gases. In general, illata exist not only in time but also in space. Reductive complexes, on the other hand, are abstracta (Reichenbach 1938, 93; Reichenbach 1951, 263). Abstracta mostly have no spatial qualities at all, although one could say that they have an existence in time. Thus the species of wallabies and a family's furniture are abstracta, as are the political state, the Bodleian Library, the spirit of the nation, and the financial crisis.

At this juncture, an important observation must be made. It concerns the so-called internal projection, a notion that will prove to be significant in Section 5, where we deal with beliefs and desires. Since illata are projective, whereas abstracta are reductive complexes, the elements of illata are of course external while the elements that constitute abstracta are internal. However, Reichenbach stresses that one and the same entity may function as an element or as a complex, depending upon the viewpoint. Thus atoms may be internal elements out of which concreta are built up, or they may be projective complexes that are inferred from concreta. In the first case the concreta actually are abstracta (they are complexes that can be completely reduced to atoms), in the second case the concreta are the (external) elements from which the atoms, as projective complexes, are probabilistically inferred. Since in the latter case the projection has a somewhat peculiar character («it leads to things which are the internal elements of the things from which the inference started»), Reichenbach calls it an internal projection (Reichenbach 1938, 216). In Section 5 we will see that Reichenbach, in the end, envisages beliefs and desires as internal projections.

2. Carnap: pure dispositions and theoretical primitives

It is interesting to see that Reichenbach's distinction between abstracta and illata has a striking parallel in Carnap's distinction between pure dispositions and theoretical constructs (Carnap 1956). As is well known, the latter distinction concerns two kinds of scientific concepts; basically it relies on the distinction between an observation language, LO, and a theoretical language, LT. Theoretical terms cannot be explicitly defined in LO and are thus introduced in LT by means of postulates. Pure disposition terms, on the other hand, occupy an intermediate position between observation terms and theoretical terms. They belong neither to LO nor to LT, but are part of a language in between the two: Carnap's extended observation language L'O.

As do the terms that denote abstracta and illata, disposition terms and theoretical terms signify non-observable or non-concrete complexes. Moreover, the criterion for distinguishing theoretical and disposition terms is the same as that by which illata are separated from abstracta. In Carnap's view, a disposition D ascribed to an object X by an investigator Y is a pure disposition if and only if there exist an S and an R such that:

(i) S is a process that affects X and is observable by Y,

(ii) R is a reaction of X and likewise observable by Y,

(iii) D is identical to (a certain combination of) S and R.

On the other hand, D is a theoretical primitive or, as I shall call it, a theoretical disposition if D is manifested by S and R, but does not coincide with S and R. Thus D is a theoretical disposition if (i) and (ii) are true whereas (iii) is false. Consequently, theoretical dispositions are only probabilistically connected to concreta. It is precisely the existence of probabilistic relations that constitutes the difference between pure and theoretical dispositions:

The decisive difference is this: on the basis of the theoretical interpretation, the result of this or of any other test or, generally, of any observations, not regarded as absolutely conclusive evidence for the state in question; it is accepted only as probabilistic evidence, hence at best as a reliable indicator, i.e., one yielding a high probability for the state. (Carnap 1956, 71; my emphasis).

Thus Reichenbach's distinction between abstracta versus illata and Carnap's distinction between pure versus theoretical dispositions are based on the same criterion. Both are grounded in the absence or the presence of probability relations. Carnap's theoretical constructs as well as Reichenbach's illata are probabilistically connected to concreta. Similarly, Carnap's pure dispositions as well as Reichenbach's abstracta coincide with sets of observable things or events. Of course, the terms signifying Reichenbach's illata and Carnap's theoretical entities are also alike: both have a surplus meaning over terms that refer to observable things or events. Thus both may be applied even if the corresponding sentences about concreta are false, or not applied even if those sentences are true. On the other hand, terms denoting abstracta or pure disposition lack this surplus, and are completely translatable into the vocabulary of observables.

Apart from the absence or presence of probability relations, there is another important similarity between the two approaches. Both the Carnapian and the Reichenbachian distinction are time- and theory-dependent. In either case, the nature of a particular non-concrete complex is not clear a priori: it has to be identified on the basis of empirical findings which in turn are based on our theory. As a result, previous decisions may be reconsidered in view of new evidence, so that, in Reichenbach's case, an illatum can become an abstractum and vice versa. The same goes for the Carnapian distinction. Time and again Carnap stressed that scientists have a certain liberty in regarding non-concrete terms as being either purely dispositional or genuinely theoretical terms; in the end, their decision is guided by considerations of empirical usefulness and efficiency, in combination with a theory at hand.

3. The question of existence

Until now we only talked about sentences and terms. We explained that the relations between sentences are either probabilistic or not, and that meanings of terms are either surplus meanings or not (all dependent upon empirical findings as well as on a theory). However, we have been reticent about the actual things in the world. To what exactly do the terms we have spoken of refer? What, if any, is the pukkah existence of non-concrete or non-observable complexes? In the present section we will address that question with respect to Reichenbach's abstracta and illata, but what we say will also apply to Carnap's pure and theoretical dispositions. As far as existence is concerned, Carnap and Reichenbach roughly held the same opinions. Both underscored that `existence' should be read as `existence-according-to-a-theory'. And both maintained that the choice of a theory has a conventional element to it (without, of course, being totally a matter of convention).

Consider again the term `species of wallabies'. According to Reichenbach this term denotes an abstractum, but does this abstractum really exist? Reichenbach's answer here is a yes-and-no. On the one hand we may say that the species of wallabies exists, meaning that many wallabies exist and that they have certain biological qualities in common which distinguish them from other animals. On the other hand, we may also say that it does not exist, meaning that many wallabies exist and that any proposition containing the term `the species of wallabies' can be translated into propositions concerning those wallabies (Reichenbach 1938, 96). For Reichenbach the question whether or not abstracta exist is settled by a decision rather than being a matter of truth-character. The decision may be an affirmation, or a denial, or neither of them. For instance, of a family's furniture we probably will say that it exists, of the height of a mountain that it does not, and in the case of human society the decision will be somewhat indeterminate. But whatever its outcome, it remains a decision and thus a matter of convention; on no account may the abstract term be taken to have a surplus meaning. The question of whether or not an abstractum exists therefore is a practical affair; regarding the matter as a theoretical topic is to raise a pseudo-problem.Foot note 5_1

Illata, on the other hand, form a different kettle of fish. Illata do have an existence of their own, and terms denoting them have a surplus meaning which goes beyond the meaning of the terms for the (external) elements. As Reichenbach phrases it: «The relation of the illata to the concreta is a projection ... The illata have, therefore, an existence of their own...» (Reichenbach 1938, 212).

The question can be illustrated on the basis of the term `atom' (cf. Reichenbach 1951). Propositions about atoms can be connected to propositions about macroscopic bodies, albeit only probabilistically: the propositions about atoms may be true whereas those concerning macroscopic bodies may be false, and vice versa. For this reason, most people will deny that the term `atom' is just an abbreviation for certain relations between macroscopic bodies. Instead, they will maintain that it refers to some thing from which those relations can be explained.

The atom example also illustrates another point that we have made above, namely that the distinction between abstracta and illata is time-dependent. The theory of the atom emerged as a pure speculation from the philosophy of Democritus in the fourth centry B.C., after which it took another twenty-two centuries before it was subjected to an empirical test. About 1800 it was found that compounds (such as for instance sucrose) consist of chemical elements (carbon, hydrogen and oxygen), of which the weights make up a fixed proportion that can be expressed in whole numbers. The English chemist Dalton realised that these fixed and quantitative relations require an explanation at the microscopic level. It turned out that all macroscopic bodies are made of microscopic particles, viz. atoms (in the case of sucrose, twelve atoms of carbon combine with twenty-two atoms of hydrogen plus twelve atoms of oxygen). By the end of the nineteenth century most philosophers and physicists agreed that atoms have an existence of their own, although there still were dissidents such as Ernst Mach, who kept believing that the word `atom' is just an umbrella term for a reducible complex.

To summarize: when taken theoretically, the question whether or not abstracta or pure dispositions exist is a standard pseudo-question. It can be answered by `yes' and `no' alike, depending on where you wish to lay the stress. If you wish to stress that sentences about abstracta resp. pure dispositions can be completely reduced to sentences about observables, then you are likely to come up with `no'. But if you wish to say that the observables in question exist as a group, then your reaction will be that abstracta do exist. All this is quite different from the illata case. There we encounter entities that do have an existence of their own, a fact that is revealed by the probability relations between sentences about illata and sentences about concreta.Foot note 5_2

4. Are there mental entities?

Beliefs, desires, and the like are non-concrete complexes of a mental kind. As such, they can be either abstracta or illata, either pure dispositions or theoretical primitives. What does Reichenbach say about them?

Reichenbach's view on beliefs and desires is quite sophisticated and certainly not the naive neo-positivistic thing that many see in it. Even scholars who sympathise with it seem to overlook how subtle and ingenious his view actually is. We could think here of Daniel Dennett, who first referred to Reichenbach's abstracta and illata in Dennett 1987. In order to get rid of the «mixed bag» of folk psychological notions, Dennett proposes «a divorce» between abstracta and illata (ibid., 57). This separation should enable us to create two tidy new theories on the mental:

one strictly abstract, idealizing, holistic, instrumentalistic -- pure intentional theory -- and the other a concrete, microtheoretical science of the actual realization of those intentional systems -- what I will call sub-personal cognitive psychology (ibid.).

By suggesting a split and directing abstracta to the one theory and illata to the other, Dennett ignores the essence of and the interesting thing about beliefs and desires, namely that they have a mixed nature. Beliefs and desires are neither plain flesh nor pure fowl. They stand «somewhere midway between abstracta and illata», and are being «pulled in two directions» (ibid., 55, 57).Foot note 5_3 Reichenbach, for his part, was fully aware of this dual nature of beliefs and desires. Rather than eliminating it by heading for a split, he tries to incorporate it by showing how abstracta and illata merge together in the mental dispositions that make up our «higher psychical life» (Reichenbach 1938, 239). Let us now see how he tries to accomplish this task.

Reichenbach's starting point is the common opinion that psychology is the science about our inner world. Next to `the higher psychical life' of beliefs and desires (with which we will deal in Section 5), our inner world entails `lower' psychic experiences such as impressions or sensations. The latter Reichenbach describes as «phenomena occurring within my mind but produced by physical things outside my mind» (Reichenbach 1938, 89-90). Examples of such phenomena are: seeing a bird flying by, feeling a man touching your elbow, hearing Donald Davidson chuckle.

Apart from the words `impressions' or `sensations', Reichenbach also uses the term `presentations' to denote the phenomena in question. What is presented by presentations are `immediate things' rather than `objective things'. Objective things are the observable things around us: tables, trees, tunes, trains. Immediate things, on the other hand, are the things experienced by the senses. The two are by no means the same. Taken as an objective thing, a pair of train rails is parallel, but taken as an immediate thing the two rails converge. Similarly, a flying bird as an objective thing differs from a seen flying bird that is an immediate thing. Objective things somehow remain the same at all times and in all spaces, whereas immediate things change with persons and perspectives: a flying bird looks smaller or larger depending on the distance, it can be taken for an airplane, a piece of paper, a drifting balloon, an UFO et cetera. Immediate things can correspond to objective things, but they can also be dreams or hallucinations. What we have called concreta are objective things; they are the physical objects or processes accessible to direct observation that form the basis of all the sciences. Immediate things, on the other hand, are the objects of psychology; they are the sort of things psychologists try to describe by referring to the basis of concreta.

As Reichenbach sees it, both the outer and the inner world can be reconstructed on the basis of concreta. This means that not only the outer world of the physical scientists, but our own inner world too can be erected on the basis of observable objects and processes. Phrased in this way, the idea is not a particularly novel one: it is shared by empiricist philosophers of all times and of all places. However, in the hands of Reichenbach this familiar thought gets an original twist. For according to Reichenbach, psychology «is a science which infers illata from concreta» (ibid., 247). To see what this means, let us take a closer look at both the concreta and the illata in question.

The concreta in psychology are observable objects or processes that can be either outside or inside your body. The outer concreta can function in two different ways, as stimuli or as responses. Typically they are stimuli whenever we are working within a first person perspective, whereas they will be responses when the third person view prevails. Thus if a car riding up causes you to believe that a car is riding up, then, from your first person perspective, the approaching car functions as a stimulus for your belief and for your subsequent action of jumping aside. The driver, on the other hand, who from his third person's perspective sees you jumping aside, may conclude that you do not want to get hit and are believing that by jumping aside you will not get hit. He describes your psychic life in terms of your reactions while you yourself are inclined to report it in terms that are taken from the stimulus sphere. Of course, the roles can be reversed: one can look upon oneself as an outsider: from the fact that you are jumping aside you yourself may safely infer that, apparently and perhaps surprisingly, you are not yet tired of life. These reversions are however exceptions. Normally one describes one's inner life by referring to stimuli, whereas the psychic life of others is mainly described by citing their observable reactions. (Mainly but not exclusively: the driver probably observed that I saw a car approaching, i.e. he noticed that a stimulus was acting upon me.)

Concreta need not be outer processes; they can also occur inside your body. In fact, Reichenbach describes two classes of inner concreta. The first is the more interesting one, since it reveals a difference between psychology and physics. It is the class of inner concreta that can only be felt by the person in whose body they occur. A physicist would certainly banish such processes as being unscientific, but in psychology they function as stimuli that are perceived only by the person who has them. An example of such an inner stimulus is the pulsation of the heart, but also a bodily awareness such as the feeling of hunger. According to Reichenbach, hunger is an inner process that is accessible to «the inner tactile sense» (Reichenbach 1938, 238). It is a concretum that is «directly observed in the same sense that we observe, say, a movement of our legs with the tactile sense» (ibid., 236). As the mentioning of leg-movement already indicates, the distinction between processes that are observed by the `inner tactile sense' and outer reactions is often not clear. Some processes, such as blushing, may be described in the reaction language as well as in the language of inner self-observation.

Concreta that function as inner stimuli differ from objects and processes that are observed by a physiologist; the latter we call inner concreta of the second class. Pictures on the retina, changes in the optic nerve, transformations of the brain, convulsions in the stomach, secretions of the salivary gland: we ourselves do not observe any of those processes. Yet they all take place in our own body, and they all can be directly observed. However, they are only observed by outsiders. Rather than being described in the language of reactions or of (inner or outer) stimuli, they are reported in the language of the physiologist who can observe the interior of bodies directly. Again, one and the same process may be described as an inner concretum of either the first or the second class, e.g. a certain process might be described as hunger or as convulsions of the stomach.

Inner concreta, whether of the first or the second class, should never be confused with illata. The confusion is easily made, since in psychology illata too are inner processes. Yet the two processes differ greatly: inner concreta can, whereas illata cannot be directly observed. Rather than being observed, illata are inferred from (inner or outer) concreta. This inference takes place along the lines of classical probability theory, since, as we have seen, between sentences about illata and sentences about concreta only probability relations exist.

What sorts of things are the illata in psychology? We have already indicated Reichenbach's surprising answer: sensations. Against the received opinion, Reichenbach argues that an optical or an acoustical sensation is not observed, but inferred. A man is exposed to an objective thing in front of him; as a consequence, he sees an immediate thing before him and has a sensation. He does not observe this sensation as he observes the thing before him or as he observes the pulsation of his heart. He must infer it, since he «does not know anything about its qualities, except that it has a certain correspondence to the immediate thing he observes. It is an unknown, X, determined as a function of the immediate thing observed» (ibid., 237).Foot note 5_4

As our explanations have hopefully shown, Reichenbach's theory of the mental cuts across the standpoints of behaviourists and non-behaviourists alike. In conformity with the habit of holding him for a logical positivist of the simple minded sort, Reichenbach is often pictured as a rigid behaviourist. It should be clear by now that this is a grave mistake. Rigid behaviourists describe people's minds in terms of their reactions to certain stimuli. They only have an eye for the outer concreta, and, since they prefer the third person view, especially focus on those outer concreta that function as reactions. Hence behaviourists have little or no interest in the essentials of Reichenbach's theory: the stimuli (especially the inner stimuli), the immediate things, the allowance of the first person view.

Needless to say, Reichenbach's theory also differs from that of the non-behaviourists, which in his case are mainly traditional psychologists fond of introspection. The proponents of introspection make the mistake mentioned above: they fail to distinguish between inner concreta and illata. If introspection means that stimuli can be inside your body (as is the case with some inner concreta), or that some things are best described by the person who has them (as is the case with immediate things), then Reichenbach has nothing against it. If, however, by introspection is meant that you can observe psychical phenomena, then you are on the wrong track. For psychical phenomena are not directly accessible by an inner sense. They are illata, that can only be indirectly inferred and never be directly observed:

The mischief of psychology does not arise from [the method of self-observation], but from a false interpretation that have been given to it. It is the concept of introspection which marks this misinterpretation, as it is meant to indicate a direct view of psychical phenomena. The interpretation developed by us, in the sense of a stimulus language, is free from such misconception ... The method of self-observation, if it is conceived as the method of stimulus language, is not less objective than reaction language. However, it opens up possibilities for observation which do not exist for the reaction method. (ibid., 243-244).

The idea of introspection is an illusion if we understand by introspection an observation of `psychical' phenomena; what we observe are physical phenomena, and the inner processes corresponding to them are only inferred. They are illata; and the basis from which we infer them is the totality of concrete objects of the physical world. (ibid., 227).

As we have seen, these `concrete objects of the physical world' exist either outside or inside our body. In the first case they are stimuli or responses, in the second case they are either stimuli or processes accessible to a physiologist. In neither case are they things that one discovers by introspection.

5. Beliefs and desires

We have seen that for Reichenbach psychology is the inference of illata from concreta, and we have explained what illata and concreta are. Our main question, however, is still unanswered. What is the nature of those entities that make up «the higher psychic life»? What are beliefs and desires? Are they the same things as sensations, viz. illata, or should we take a different view? It is time to deal with these issues, and to explain why Reichenbach's view is so interesting for us today.

Like Carnap, Reichenbach regards beliefs and desires as dispositions. And as we have explained, both Reichenbach and Carnap hold that there are two sorts of dispositions: Carnap distinguished between pure and theoretical dispositions, while Reichenbach argued that dispositions can be either abstracta or illata. However, Reichenbach claimed that psychological dispositions such as beliefs and desires are always abstracta. What is more, they are abstracta of a special kind. For the internal elements that compose beliefs and desires are not only concreta (as in ordinary abstracta) but also illata:

Psychology is a science which infers illata from concrete objects. The inferred objects are projective complexes of these concrete objects. Since some of the objects of psychology such as bodily feelings are accessible to the inner tactile sense, the inferred illata in such cases are internal elements of the observed concrete objects; it is therefore the process of internal projection which plays a role here. The `higher' psychological objects, and just those most frequently occurring in practical psychology, i.e., psychology as needed for daily life, are abstracta, built up of concreta and illata. (ibid., 1938, 247).

Thus the situation seems to be as follows. In psychology we aim at knowledge of people's minds, including our own mind. Since we are unable to read minds directly, we must start by looking at people's bodies, including our own body. What we then see is a number of objective things: a motorcar approaching rapidly, a man next to us jumping aside, a sharp pain in our left arm, a bone sticking out of the man's leg. The example is not very pleasant, and perhaps I should apologise for that, but it illustrates clearly the four objective things that we have distinguished: outer stimuli, outer responses, inner stimuli, and the objects of physiological observation.

Of course, I am the one who senses these objective things, and therefore a number of immediate things is present too. There is a motorcar and a seen motorcar, a jumping man and a seen jumping man, a bone and a seen bone. This does not mean, however, that objective and immediate things coincide. The psychologist who afterwards is going to treat me for the trauma caused by witnessing this accident is not interested in the objective events. She primarily wants to know what went on in my mind. That is, she wants to know what my impressions are, and those are characterised by the immediate rather than by the objective things. The objective events are only important in so far as they teach us something about my impressions. At the same time, however, objective things are the only things my psychologist can rely on. Nothing else than physical concreta can guide her -- or me -- in the search for what exactly went on in my mind. But since physical concreta do not fully determine my mental impressions, we need probability relations to infer the immediate from the objective things.

And now we are able to draw an interesting conclusion. In Reichenbach's view, the sensations, feelings, impressions or `presentations' that make up the `lower' psychic life are all illata: they are immediate things that are inferred from concreta. Together with (other) concreta, these illata constitute our `higher' psychic life, notably our beliefs and desires. From this it follows that beliefs and desires are abstracta. However, they are abstracta of a special kind. For they are composed not just of concreta, but of concreta and illata. Hence the illata in question, that were originally inferred from concreta, now function as internal elements of abstracta, c.q. of beliefs and desires (cf. Section 1). This means that beliefs and desires have a mixed nature. They are neither pure abstracta nor pure illata, neither plain reductions nor pure projections. Rather they are examples of what in Section 1 were called internal projections.

Those are the outlines of Reichenbach's theory of the mental. What lessons can be drawn from them?

The first lesson concerns the assumption, also entertained by Carnap, that the mental and the physical are related in a probabilistic way. Reichenbach developed this assumption in his notion of probability meaning, whereas Carnap made it the basis for his theories of inductive logic. Neither of the two projects proved to be entirely successful, but that does not mean that research into probability connections is doomed to disappointment. After all, the idea that probabilistic features play a role in the relation between mental and physical features is by no means unrealistic. It is therefore somewhat surprising that the notion of probability seems to be entirely forgotten whenever one talks about relations of emergence or supervenience. Elsewhere we have written about probability (Atkinson & Peijnenburg 1999), and we will not dwell upon the subject here.

The second lesson pertains to Reichenbach's explanation of the first person view. Self-observation has always been a problem for empiricists, naturalists, physicalists and, in general, all philosophers who are scientifically oriented. On the one hand they cherish the idea that outer, verifiable events make up the basis for science, and, in fact, for any meaningful statement. On the other hand, they deplore an all too rigid approach, in which any special access of a person to (part of) his mental life is bluntly denied. Reichenbach offers a way out. He runs with the hare and hunts with the hounds by making a distinction between inner concreta (that can be self-observed) and illata (that are only inferred). In this manner introspection becomes a fact, but it is no more extraordinary than the fact that I can see your back whereas you cannot. Of course, Reichenbach is not the only one who offered an empiricist solution for the problem of first person authority. It cannot be denied, however, that his approach is rather original and worthy of further exploration. In any case it is a welcome antidote for the increasing number of approaches that have abandoned empiricism altogether (cf. the transcendentalism of White in White 1991).

The third lesson is the most important one. Simply put, it boils down to the advice that we should be tolerant. By this I do not mean Carnap's famous adage about freedom in the choice of language systems («Let us be cautious in making assertions and critical in examining them, but tolerant in permitting linguistic forms»), although Reichenbach no doubt would have heartily endorsed this maxim. What I mean is that we should not try to fit beliefs and desires into the straitjackets of being either an ordinary abstractum consisting of concreta or an illatum inferred from concrete events. Beliefs and desires have a dual character, which Reichenbach tries to acknowledge by saying that they consist of concreta and illata alike. We have seen that Dennett too notices this dual character (Dennett 1987); however, rather than broad-mindedly accepting this fact, he tries to get rid of it by suggesting a divorce, culminating in two totally different theories, one about abstracta and one about illata.

Once the dual nature of beliefs and desires is taken seriously, some strong hunches are easily accounted for. For example, it is highly unlikely that all beliefs and desires are either abstracta or illata. It is much more plausible that they, like all dispositions that make up the higher psychic life, exhibit gradual differences. Thus some will be very close to pure abstracta, while others depend for the greater part on theoretical entities or illata. It seems only natural to regard for instance politeness and prosperity as abstracta or pure dispositions: it is unlikely that they will ever be more than abbreviations for a cluster of responses which appear under certain circumstances. Aggressivity and claustrophobia, on the other hand, presumably are illata. It is quite possible that future research will find that frequent aggressive behaviour corresponds to sensations caused by a chemical substance or a physical entity (the pugnacity lobule? the truculence particle?). By taking a tolerant stance and accepting that the nature of beliefs and desires is mixed, we can make these intuitions plausible. Hence we can avoid an all too monolithic approach to the higher psychic life, and learn to see reliefs in the mental map.

The old empiricists divided the mental into impressions and ideas. Modern empiricists have adopted this division by distinguishing between feelings or sensations on the one hand and pro-attitudes on the other. The criteria for the division correspond to two major themes in the contemporary philosophy of mind, viz. consciousness and content. According to almost everybody in this field, impressions or sensations are things of which we are conscious or aware; they are characterised by qualia. Ideas and pro-attitudes, on the other hand, are said to have content; they are characterised by intentionality or `aboutness'. In general, consciousness is seen as the fundamental phenomenon, upon which intentionality ultimately depends. There are however dissident philosophers, such as Dennett, who think that the order should be reversed. Be that as it may, the two great problems in the philosophy of mind are exactly about these two features: how to give an account of qualia and what exactly is intentionality? Reichenbach's position, old though it may be, might shed new light upon both of them. For as we have seen, Reichenbach regards sensations, impressions and thus qualia, too, as things that are probabilistically inferred rather than directly felt. Moreover, he conceives pro-attitudes as being composed of concreta and illata, thus making it more easy to understand that beliefs and desires have a mixed nature, and that some are `more real' than others.Foot note 5_5


Atkinson, D. and J. Peijnenburg (1999), `Probability as a Theory Dependent Concept', Synthese, 118, 207-228.

Carnap, Rudolf, (1936-1937), `Testability and Meaning', Philosophy of Science 3(4), 1936, 420-471, and 4(1), 1937, 1-40.

--- (1950), `Empiricism, Semantics, and Ontology', Revue Internationale de Philosophie 4, 20-40. Reprinted in: Carnap 1947, 205-221.

--- (1947), Meaning and Necessity. A Study in Semantics and Modal Logic. Chicago: University of Chicago Press, 1947. Enlarged edition 1956, second impression 1958.

--- (1956), `The Methodological Character of Theoretical Concepts', in: H. Feigl and M. Scriven (eds), The Foundations of Science and the Concepts of Psychology and Psychoanalysis. Minneapolis: University of Minnesota Press, 38-76.

--- (1963), `Intellectual Autobiography', in: Schilpp, P.A. (ed.), The Philosophy of Rudolf Carnap. La Salle, Ill.: Open Court; London/Cambridge: Cambridge University Press, 1963, 3-87.

--- (1966), Scheinprobleme in der Philosophie: das Fremdpsychische und der Realismusstreit. Frankfurt a. M.: Suhrkamp.

Dennett, Daniel (1987), The Intentional Stance, Cambridge (Mass.), MIT Press.

--- (1991a), Consciousness Explained, Boston: Little Brown. Published in 1993 by Penguin Books, Harmondsworth, Middlesex, England.

--- (1991b), `Real Patterns', Journal of Philosophy 89, 27-51.

Peijnenburg, Jeanne, and Ronald Hünneman, `Translations and Theories', to appear in Ratio, March 2001.

Reichenbach, Hans (1935), Wahrscheinlichkeitslehre: eine Untersuchung über die logischen und mathematischen Grundlagen der Wahrscheinlichkeitsrechnung. Leiden: Sijthoff.

--- (1938), Experience and Prediction. An Analysis of the Foundations and the Structure of Knowledge. Chicago: The University of Chicago Press, seventh impression, 1970.

--- (1951), `Are There Atoms?', in: The Rise of Scientic Philosophy. Berkeley: University of California Press, Chapter 11.

Smullyan, Raymond (1977), `Is God a Taoist?', in: The Tao is Silent, Harper & Row. Reprinted in: Douglas R. Hofstadter and Daniel C. Dennett (composers and arrangers), The Mind's I, New York: Basic Books, 1981, 321-341.

White, S. (1991), The Unity of the Self. Cambridge, MA: MIT Press.

Jeanne Peijnenburg

University of Groningen

A-weg 30

9718 CW Groningen. The Netherlands