1 Introduction
The doctrine of physicalism in the Vienna Circle, according to Feiglâs retrospective formulation, comprised of two theses:
- i.Physicalistic language1 is the universal language of science.
- ii.All scientific laws and facts can be reduced in principle to the fundamental laws and facts of physics.2
And we may add:
- iii.No ontological thesis about the nature of the world had been advocated, according to which âEverything is physicalâ or âThere is nothing over and above the physicalâ, or âEverything supervenes on the physicalâ,3 as held by later-day metaphysical physicalists.4
The language thesis (i) was generally accepted in the 1930s by the physicalists of the Vienna Circle, while the nomological reducibility thesis (ii) was disputed. Carnap, for example, regarded it as a possible but as yet unjustified hypothesis; Neurath, however, was against such a reductionist conception at least from 1932 (see his 1932a),5 and promoted his own distinct conception
2 Part 1: Carnap
2.1 Carnapâs Views on Autopsychological Sentences
1930 and early 1931: Proto-Universalsprache and Proto-Psychologie. In the beginning of his physicalist period, between 1930 and 1932, Carnap changed his account of autopsychological sentences several times. In 1930 he was already endorsing a universal language doctrine, according to which all scientific statements can be formulated in physicalistic language, including all psychological statements. He gave several lectures promoting this view,6 and in 1930 he had already written the first versions of his papers later to be published as âDie physikalische Sprache als Universalsprache der Wissenschaftâ and âPsychologie in physikalischer Spracheâ.7 According to these draft papers, dubbed âproto-Universalspracheâ and âproto-Psychologieâ by Thomas Uebel,8 system statements about psychological states must be formulated in physicalistic language, that is, about bodily and behavioral states of persons, while protocol sentences, describing the observations of protocollers verifying the system sentences, ought to be formulated in phenomenalistic language. The reason for Carnapâs adherence to phenomenalistic protocol sentences was his assumption that judgements about autopsychological experiences were
1932. Universalsprache and Psychologie. After substantially revising the drafts, Carnap published âUniversalspracheâ and âPsychologieâ in May and December 1932. In these papers he elaborated the thesis that all scientific sentences should be formulated in physicalistic language in detail. According to his new proposal, not only system statements but also protocol statements should be expressed in physicalistic language.
Nevertheless, in âUniversalspracheâ Carnap also addressed the question of how to analyse the content of perceptual experiences, i.e. perceptual reports formulated in phenomenalistic language. He suggested that phenomenalistic sentences expressing the content of perceptions may have different logical forms: they may refer to atomistic sense data, or to Gestalts or to physical objects, i.e. to medium-size macroscopic material objects and their properties. Carnap did not choose between these options in the paper.
The reason why Carnap discussed this topic at all is somewhat puzzling, considering that his new account required protocol sentences to be physicalistic. This apparent contradiction or tension may be resolved, however, by noting that Carnap required only perceptual reports serving as protocol statements to be expressed in physicalistic language. For unless they are physicalistic, they could not be understood intersubjectively, hence they were inapt for scientific purposes as they could not verify system statements. But this does not entail that all phenomenalistic sentences are meaningless, or that solipsistic phenomenalistic languages per se are unintelligible (though, of course, phenomenalistic languages cannot be part of the universal language of science).
Moreover, according to Carnap, phenomenalistic sentences about experiences with a private meaning can be translated into physicalistic sentences that have a public, intersubjective meaning. For example, the private phenomenalistic sentence p âRed (is being seen by S) nowâ is to be replaced with the physicalistic sentence âThe body S is seeing red nowâ to be understood as âS(âs body) is in a red-seeing state nowâ.9 Expounding on what such a translation entails may reduce the prima facie implausibility of the suggestion. First, Carnapâs explanation of the claim that the protocol statement p âRed (is being seen by S) nowâ, can be translated into the physical statement P âSâs body is in a red-seeing state nowâ is that the scientific content of a sentence is exhausted
It is, of course, intuitive that we mean something different by the statements âRed (is being seen by S) nowâ and âSâs body is in a red-seeing state nowâ (the former also including that S has a red experience beyond being in a certain bodily state). But this, according to Carnap, is mistaken. The illusion that âRed (is being seen by S) nowâ has a surplus meaning over âSâs body is in a red-seeing state nowâ stems from confusing the logical content of the sentence âRed (is being seen by S) nowâ (which is: Sâs body is in a red-seeing state now) with the accompanying image, the content of the accompanying representation (Vorstellungsgehalt), (i.e. that S has a red experience).11
Second, it should also be noted that since the physicalistic sentence P involves no reference to Sâs experience, prima facie, the possibility of such a translation does not require that there be a general parallelism between Sâs experiences and bodily states/behavioural dispositions. It is enough if a diachronically constant correlation between certain physical stimuli and bodily states and behavioural responses of S exists. Thus, the availability of such physicalistic translations of psychological sentences does not depend on the capacity to know other personsâ experiences â which would be necessary, if the possibility of translation required the content of psychological sentences to be identified by reference to the experiences of their owners and then paired with the physicalistic sentences, referring the bodily states/behaviour correlated with the experience with which they are to be substituted.12
1932. Ãber Protokollsätze. In âÃber Protokollsätzeâ (1932b/1987),13 which was a reply to a criticism of Carnapâs account (above) put forth by Neurath in
The objection from different meaning. Phenomenalistic protocol sentences seem to have a different meaning than physicalistic protocol sentences, that is, they are not only reporting about bodily states, but also about the perceptual experiences of the observer. Carnapâs answer was the same as in âUniversalspracheâ, namely that the objection rests on confusing the factual content of protocol sentences (the physical state of the protocollerâs body when in a certain perceptual state), with the accompanying representation of that state (i.e., the phenomenal experiences accompanying the bodily state).
The objection from the individuation of psychological states. Psychological states are private in the sense that they are presented only to their subject. How is it then possible to individuate psychological states by physicalistic sentences (i.e. by reports about physical states) the meaning of which is intersubjective?
My perception of red, my hunger, my anger are presented only to me but not to my neighbours. However, in physicalism all sentences are intersubjective. Where do we locate the fact about the immutable separation of subjects from each other? This fact should not be denied; but it must be carefully formulated. âS is hungryâ is synonymous with âThe nervous system of S is in a hunger stateâ; âS sees redâ is synonymous with âThe nervous system of S in a red-sensing stateâ. âOnly S is immediately aware of his hungerâ means: âOnly S is able to make the statement âS is hungryâ â directly based on his hunger, i.e., with no physical causal connection with processes outside of Sâs bodyâ.14
Let us analyse Carnapâs proposal for individuating psychological states in some detail.
First of all, it should be noted that Carnapâs aim was not to answer the question âhow is it possible that physicalistic sentences that have an intersubjective meaning may refer to private experiences?â Such an objection would be
Thus, Carnap was after something different. He aimed to explain how psychological states can be individuated and discriminated from each other using physicalistic descriptions, by referring to bodily states and behaviour alone. It is not obvious first how to do this. In contrast, individuating psychological states using phenomenalistic terms is quite straightforward. Since, by assumption, each particular experience may belong only to one person, any experience can be individuated by reference to the person who has it. Persons, in turn, could be identified by a set of compresent experiences and by a personal history, i.e. a temporal series of experiences connected by causal and memory relations between them.15 However, if reports of psychological states cannot refer to experiences, only to bodily states and behaviour, as physicalism suggests, then this way of individuating persons is not available.
Why not, then, identify a psychological state by reference to the bodily state/behaviour it is correlated with? Unfortunately, this is also unavailable, since it presupposes that we directly identify another personâs experience and pair it with the bodily state/behaviour with which it is co-instantiated. But such identification of othersâ experiences is not possible.
Carnapâs suggestion is as follows. We identify a psychological state, for example, a particular state of hunger, by reference to the neural state which directly causes (i.e., without involving inference) the verbal utterance âHunger nowâ displayed by Sâs body (mouth), such that the causal process leading from this neural state to the verbal act of reporting takes place exclusively within Sâs body. This way, a particular psychological state can be identified by reference to that particular bodily state which directly causes the utterance âHunger nowâ without referring to the experience of hunger. This means that the psychological state can be identified by relying only on physicalistic sentences.
Furthermore, the particular psychological state can be distinguished from other similar psychological states (e.g. from other states of hunger instantiated by some other person at the same time). For S may utter the same sentence âHunger nowâ, reporting a psychological state upon perceiving S*âs, another personâs bodily state and behaviour characteristic of being hungry. In such a
It is important to see that Carnapâs proposal for individuating psychological states still does not undermine the view that phenomenalistic autopsychological sentences are meaningful. For while it is true that one cannot formulate the criterion that distinguishes between persons in phenomenalistic language, this, prima facie at least, does not exclude that the use of phenomenalistic terms by a particular person referring to his or her own experiences may be constant.
2.2 Carnapâs Implicit Views about the Nature of Experience Supporting a Phenomenalistic Account of Autopsychological Sentences
After presenting Carnapâs account of autopsychological sentences, I will now discuss some views about the nature of experience and experiential states, which I believe form the background of this account. In my opinion, Carnap was committed to the following theses.
- (1)Experience is logically private.16
- (2)The individual Given has a diachronically constant nature. In other words, in similar environments and under similar perceptual conditions a personâs experiences at different times are phenomenally similar.
- (3)We have reliable direct/introspective capacities to diachronically re-identify our experiences.17
An argument for the logical privacy of experience, consonant with Carnapâs physicalism, is that a particular experience is related only to one body (i.e., to the experiencerâs âown bodyâ). More precisely: some bodily sensations, at least, are localized in some particular body part: for example, a certain pain is a tooth-ache or a head-ache etc. Since a body part cannot belong to more than one body (except in conjoined twins, but letâs put this aside), a particular bodily sensation can belong only to one body.
A plausible objection against such an account is that it is at least logically or conceptually possible that a particular experience of a person correlates with the physical state of some other personâs body, not his own, and in such cases it is unclear whether the experience can be individuated by reference to the body the stimulation of which gives rises to the experience, i.e. to the other personâs body, if the other person also has an experience of the same kind. Carnap discussed such cases as well. In âPsychologieâ, he considered the (hypothetical)
However, in his reply to the objection from telepathy,19 Carnap seems simply to assume that experience is private but does not argue for this claim. Nonetheless, there are arguments which aim to establish that it is conceptually impossible to interpret such cases of telepathy as one person being aware of some other personâs experiences. Or, in other words, that it is logically impossible to feel an instance of fear which is numerically identical with some other personâs fear. Wittgenstein, for example, between 1929 and 1932, proposed on several occasions that the statement âI feel your painâ is nonsensical and logically impossible. Wittgensteinâs main points were that accepting such a description of the situation when my feeling of pain is correlated with another personâs bodily pain state would commit one to the â according to him, absurd â assumption that there exist experiences no-one ever has; and moreover, that such a view rests on a false analogy between introspection and the perception of physical objects.20,21
As for the arguments for (2) and (3), Carnap seems to take them for granted. One strong piece of evidence for this is that in the Aufbau, the construction system is based on the recollection of similarity, and the capacity to determine the similarity between present and past experiences is direct. One could argue that referring to Carnapâs Aufbau-views, in the proposed context, is misguided, since he had already moved away from these views in the physicalist period I am investigating. This is, of course, true of many aspects of the Aufbau. Most importantly, Carnap gave up the whole idea that the system of scientific concepts should be logically constructed on a methodological solipsist phenomenalist base. But this notwithstanding, there is no indication in the papers
3 Part 2: Neurath
According to the received view in Vienna Circle scholarship, Neurath had a major influence on Carnapâs physicalistic turn. This influence has been discussed in great detail by Thomas Uebel in many of his writings,23 and also by Juha Manninen.24 I shall strongly rely on these investigations as a background to the issues I discuss in the remaining sections of the paper. In what follows, after setting the context regarding the changes of Neurathâs accounts and arguments for physicalism between 1929 and 1932, I will concentrate mainly on what may be called âNeurathâs private language argumentâ. I shall analyse this argument in some detail, compare it with Carnapâs versions of physicalism, and also discuss whether Neurathâs argument may have influenced Carnap. My claim is not that Neurathâs multifarious efforts to promote physicalism did not play an important role in Carnapâs physicalistic turn. Rather, I would suggest more modestly that Neurathâs âprivate language argumentâ was not accepted by Carnap, and hence that particular idea of Neurath did not contribute to Carnapâs opting for linguistic physicalism in the early thirties.
3.1 Neurathâs Views on Autopsychological Sentences
Neurathâs commitment to physicalism emerged much earlier than Carnapâs, in the first half of the 1920s (partly originating from historical materialism).25
From 1931 onwards, however (at meetings of the Vienna Circle in early 1931, in his papers âPhysikalismusâ (1931), âSoziologie im Physikalismusâ (1931) and âProtokollsätzeâ (1932)),28 Neurath changed his mind, and argued for a more radical physicalism, featuring the claim that private phenomenalistic languages are not only inapt for scientific purposes, but also meaningless. He claimed not only that protocol sentences cannot be formulated in phenomenalistic language, but also that solipsistic phenomenalistic languages â created and used only by one person â are impossible. In the following, I reconstruct Neurathâs main points in favour of this view.
3.2 Neurathâs âPrivate Language Argumentâ
If someone makes predictions and wants to check them himself, he must count on changes in the system of his senses, he must use clocks and rulers; in short, the man who supposedly is in isolation already makes use of the intersensual and âintersubjectiveâ language. The forecaster of yesterday and the controller of today are, so to speak, two persons.29
The exclusion of âphenomenal languageâ in its present form, which does not seem to be even suitable for âpredictionsâ, that is, for what is essential for science, will probably necessitate a number of alterations in the constitutional (constitutive) system. Together with this the âmethodological solipsismâ (Carnap, Driesch) will probably also disappear.30
This was indeed a private language argument. Against Carnapâs methodologically solipsist protocol language, Neurath argued that a language must be usable by one individual over time, and he derived from this [â¦] the condition that there be âconstancy of useâ.
[â¦]
Neurath did not argue that phenomenal language were logically impossible, Carnap had obviously shown how to construct one. Neurath nevertheless contended that [â¦] the belief that oneâs language use is constant, must be justifiable. But it is not, if phenomenalist languages are being spoken. With intersubjective language, by contrast, it is reasonable to claim that had oneâs use become inconsistent as failed attempts at communication would have become ever more frequent and so the inconstancy would have shown up. Since they did not, oneâs language use did remain constant.
[â¦]
Neurathâs new argument (i.e., the private language argument â g.a.), presupposed the view that only physicalistic language provides for public references whose accessibility makes for intersubjective language. (â¦) Neurath noted about phenomenal language that âit does not seem to be even suitable for âpredictionsâ that is, for what is essential for scienceâ. [â¦] his meaning was clearly that predictions formulated in the phenomenal language could not be tested by others. What was new [â¦] that he added an argument which focussed on what was required for each individual agent to comprehend and systematize their own experience (It did not focus only on what was required for communication.)31
- (a)intersubjectively, by others,
or
- (b)subjectively, by the subject him or herself.
As for (a), the justification of intersubjective unverifiability: although the use of the terms of a private phenomenalistic language by a subject may be diachronically constant, this cannot be verified by others, because of the logical privacy of experience. In order to make this point more clearly, it is illuminating to consult Wittgensteinâs discussion of the âbeetle in the boxâ in Philosophical Investigations, sec. 293. According to Wittgenstein, it is conceivable that different persons have different sort of beetles in their respective boxes, or that a beetle in a box is constantly changing over time, or even that there is no beetle in someoneâs box at all. If the use of the world âbeetleâ is constant across the different language users, then the word âbeetleâ cannot refer to a private object.32 Neurath holds, similarly, that because of the logical privacy of experience, psychological reports of a person cannot be about his or her experience, for then the constancy of his or her use of phenomenal terms could not be checked by others. Therefore, a personâs predictions (which are of fundamental importance for science) could not be checked by others either.
As for (b), the justification of subjective unverifiability: according to Uebel
Neurath held that even a solitary thinker required a system of symbolic representation for the ordering of his experience over time which was of necessity intersensual and intersubjective. A phenomenal language does not âcome into questionâ for it does not allow for the mechanisms whereby the constancy of an individualâs language use can be controlled by the individual himself allowing predictions to be checked.
And
Once on a solipsistic base, there is no preventing solipsism-of-the- moment.33
In my view, we may distinguish two strands (not completely independent of each other) in Neurathâs reasoning. According to the first, which may be called the argument from verification, the constancy of use of phenomenal terms cannot be verified subjectively, because there is no means or method for determining whether memory representations are correct, if the subject can only rely on what is introspectively accessible to her; i.e. if the only test of the subjectâs memories being correct is that they seem to be so to the subject. According to the second, what we may call the argument from symmetry, a personâs predictions about the perceptions of her âfuture selfâ (i.e. a diachronically distinct person, who is conveived as identical with her),34 have the same epistemic status as a personâs predictions about the experience of another synchronically existing distinct person. The first point is aptly expressed by Uebelâs formulation that solipsism leads to solipsism-of-the-moment, and the second by Neurathâs formulation that the forecaster of yesterday and the controller of today are, so to speak, two persons.
3.2.1 The Argument from Verification in Detail
According to Neurath, the controller cannot know whether her experience is of the same sort the forecaster predicted, because her senses may change. It is possible that an experience seems to the controller to be an instance of the sort the forecaster predicted, but in fact, it is not; and also, it may seem different to the controller, while it is, in fact, the same. If the only way for the controller to test whether her current experience is of the sort the forecaster predicted that she (the controller) would have, is to rely on that it seems to the controller to be so (in other words, if the controller can base her judgement only on that it seems to her that her current experience is the same sort she remembers as appearing in the forecaster prediction), then the controller does not know whether her experience is of the same sort the forecaster predicted.
The source of the problem is not that memory may not be reliable. It is possible that the memory of the controller is reliable, thus, when the controller
In order to the controller to know that she is observing what the forecaster predicted, she must be able to identify the content of her observation objectively/intersubjectively. She cannot do this by relying solely on the phenomenal content of her perceptual experience introspectively accessed, but only by identifying the external physical objects of her perception, which can be executed only by objective/intersubjective means, by âclocks and rulersâ, that is, by the space-time localization of the object of perception. These instruments and the space-time localization, however, already involve reference to public, intersubjective objects, so the content of perceptual experience is already identified by intersubjective notions and language. So, Neurathâs fundamental point is that without relying on external means, the constancy of the meaning of terms cannot be guaranteed, which is tantamount to the claim that even a private, subjective phenomenal language used by a âsolitaryâ thinker, serving only for personal use, is impossible.
258. Let us imagine the following case. I want to keep a diary about the recurrence of a certain sensation. To this end I associate it with the sign âSâ and write this sign in a calendar for every day on which I have the sensation. â I will remark first of all that a definition of the sign cannot be formulated. â But still I can give myself a kind of ostensive definition. â How? Can I point to the sensation? Not in the ordinary sense. But I speak, or write the sign down, and at the same time I concentrate my attention on the sensation â and so, as it were, point to it inwardly. â But what is this ceremony for? For that is all it seems to be! A definition surely serves to establish the meaning of a sign. â Well, that is done precisely by the concentrating of my attention; for in this way I impress on myself the connexion between the sign and the sensation. â But âI impress it on myselfâ can only mean: this process brings it about that I remember the connexion right in the future. But in the present case I have no criterion of
correctness. One would like to say: whatever is going to seem right to me is right. And that only means that here we canât talk about ârightâ. 265. Let us imagine a table (something like a dictionary) that exists only in our imagination. A dictionary can be used to justify the translation of a word X by a word Y. But are we also to call it a justification if such a table is to be looked up only in the imagination? â âWell, yes; then it is a subjective justification.â â But justification consists in appealing to something independent. â âBut surely I can appeal from one memory to another. For example, I donât know if I have remembered the time of departure of a train right and to check it I call to mind how a page of the time-table looked. Isnât it the same here?â â No; for this process has got to produce a memory which is actually correct. If the mental image of the timetable could not itself be tested for correctness, how could it confirm the correctness of the first memory? (As if someone were to buy several copies of the morning paper to assure himself that what it said was true.) Looking up a table in the imagination is no more looking up a table than the image of the result of an imagined experiment is the result of an experiment.35
We may briefly reconstruct Wittgensteinâs argument as follows.
- (1)The condition for a term having meaning is that there must be a verifiable difference between applying the term correctly and incorrectly.
- (2)If there is no verifiable difference between a present experience being of the same sort as some earlier experience S, and it is only seeming as if the present experience were of the same sort as S, then the sensation term âSâ, putatively referring to S, has no meaning.
- (3)If one can rely only on introspection, then one cannot determine whether a present experience is of the same sort as some earlier experience S, or whether it only seems so.
Hence
- (4)There is no verifiable difference between using the sensation word âSâ correctly or incorrectly.
Hence
- (5)âSâ has no meaning.
Neurathâs point seems similar. Neurath claimed that if a person wants to determine whether her present experience is of the same sort as some earlier one, and she can rely only on introspecting her experiences, then she has âno
Now, the moral of Wittgensteinâs argumentation in Philosophical Investigations 243 and 258 is similar: to determine whether the use of a word is constant through time (i.e. whether we take it to refer to the same object) is possible only if the word refers to some public, intersubjectively accessible object. Conjuring up memory images of train timetables, in itself, cannot justify my belief about the departures and arrivals of trains. Such memory images may be justificatory only if the content of the timetables is grounded in the objective facts regarding the departures and arrivals of trains. Without such grounding, memories of the timetables cannot establish that I have remembered the departure and arrival times of the trains correctly. Similarly, trying to justify the belief that my current experience is of the same sort as a particular past experience of mine by trying to remember or to imagine other instances of the putatively same experience is useless. Remembering or imagining instances of experiences may be justificatory only if there are some external criteria determining whether in these cases of remembering or imagining, the subject is in the same internal, i.e., experiential, state or not.
3.2.2 The Argument from Symmetry in Detail
What counts in favour of Neurathâs suggestion that the controller is distinct from the forecaster in the same way as two synchronously existing persons are distinct from each other, as concerns their access to the otherâs mind?
In my view, the point Neurath wants to make is that the controller has no direct access to the sensations or perceptual experiences of the forecaster (i.e. her âpast selfâ), just in the same way as a particular person has no access to the experiences of other synchronously existing persons. A person may have assumptions about another distinct personâs experiences based on the observational situation the other person is in, and she may have imaginative representations of the other personâs experiences. But these, of course, do not provide direct access to the other personâs mind, to her experiences. Similarly, the controller is not in the position to know, based only on her introspectively accessible experiences and memories, whether the experience the forecaster predicted is the same kind of experience the controller is now having. For
3.3 Neurathâs Implicit Views on the Nature of Experience Supporting the Physicalistic Account of Autopsychological Sentences
In my opinion, Neurathâs implicit account of the nature of experience and conscious states37 backing his private language argument may be characterized as follows.
i.e. the phenomenal content of experience, may not have a diachronically constant nature (neither in its monadic nor in
- (3)We do not have a reliable direct, introspective capacity to diachronically re-identify our experiences.
In my understanding, both Neurathâs argument from verification and his argument from symmetry rely on (2) and (3). (3) is more fundamental; if we do not have a reliable capacity to re-identify diachronic experience, then the claim that the phenomenal content of experience does not have a diachronically constant nature cannot be true and since it is unverifiable, it has no truth-value.40
Clearly, if we had such a capacity (and we could verify whether the phenomenal content of our experiences is diachronically constant or not), then both of Neurathâs arguments could be rejected. On the one hand, it would be false to claim that one cannot subjectively verify that a particular present experience is of the same sort as some past experience. If I had a reliable capacity to re-identify diachronic experience then the fact that it phenomenologically seems to me that the present experience is like the earlier experience would indeed justify my judgement (its seeming to me in the judgemental sense) that the present experience is of the same sort as the earlier experience. Such a verification is not conclusive, of course. Judgements based on such a capacity to compare and determine the identity of qualities of experiences at different times are not infallible, but they may be reliable.
Furthermore, if we had a reliable capacity to re-identify diachronic experience, then the symmetry thesis would not stand either. For then, in the diachronic case, there existed a reliable epistemic mechanism that would ground the controllerâs judgement about the completion of the forecasterâs prediction; and the controller would be justified in her judgement about the phenomenal similarity of her present experience to the past experience of the forecaster. At the same time, in the synchronous case, a personâs imaginative representation would not provide such a reliable epistemic mechanism for obtaining knowledge about another personâs experience. Hence, such judgements about other personsâ experience would not have the same epistemic status as the controllerâs judgements about the completion of the forecasterâs predictions.41
4
In 1932, in âUniversalspracheâ (Carnap 1931/1934), in âPsychologieâ (Carnap 1932a/1959)and in âÃber Protokollsätzeâ (Carnap 1932b/1987), Carnap suggested that protocol sentences should be formulated in physicalistic language. But he did not claim in these papers that phenomenalistic languages are unintelligible. Consequently, Carnap did not change his mind, and suggested that protocol sentences should be formulated in physicalistic language under the influence of Neurathâs private language argument. To this extent my view may be in conflict with the generally accepted view that Neurath was the major source of Carnapâs physicalist turn. I, however, do not dispute Neurathâs having an important role in turning Carnap to physicalism, only that Neurathâs private language argument was not among its motivating factors, for, as I see it, Carnap did not accept it.
It seems though, that Carnap did not explicitly reject Neurathâs private language argument, but rather ignored it, deeming it not so much as an argument but a bundle of unelaborated ideas, apercus. Nonetheless, we may investigate what Carnap could have answered to Neurath. All the more so, since it seems obvious that Carnap must not have accepted Neurathâs private language argument even later, as he maintained that private phenomenalistic languages are intelligible (as late as in 1961, in his introduction to the English edition of the Aufbau, and in 1963 in his reply to Goodman).42, 43 So, could Carnap rebut Neurathâs private language argument based on his views?
However, in my view, Carnap assumed in his early physicalist period, between 1930 and 1932, that we have a reliable introspective capacity to diachronically re-identify our experiences. (And probably also later; cf. footnote 43). If this is the case, then both the claim that one cannot subjectively verify the constancy of use of her phenomenal terms, and the claim that synchronically and diachronically distinct persons are in a symmetrical position to know each otherâs experiences, can be rejected, as argued above. Moreover, there is a bonus point: if we assume that we have a capacity to re-identify experience diachronically, then it seems that the verificationist interpretation of
I use âphysicalistic languageâ to refer to languages the terms of which denote ordinary physical objects and properties, i.e. observable entities and properties, and âphenomenalistic languageâ which refer to conscious experiences. Hence, physicalistic language is not the language of physics; its terms do not refer to entities or properties postulated by physical theories, as e.g. âelectronâ or âchargeâ. âPhysicalâ and âpsychologicalâ refer to objects or states which are commonsensically regarded as physical (bodies and planets, being heavy or solid), and as mental (persons, beliefs, desires, perceptions), respectively. Hence, we may use the following formulations: according to linguistic physicalism, reports about both physical and psychological phenomena are to be couched in physicalistic language, according to linguistic phenomenalism in phenomenalistic language.
Cf. Feigl 1963, pp. 227â228.
For a summary about the contemporary formulations of metaphysical physicalism, see Stoljar 2017.
Thus, in line with the general anti-metaphysical attitude of the logical empiricists, Carnapâs and Neurathâs physicalism does not mean the reduction of âeverythingâ, including mental phenomena, to physics.
âThe development of physicalistic sociology does not mean the transfer of laws of physics to living things and their groups, as some have thought possible. Comprehensive sociological laws can be found, as well as laws for definite narrower social areas, without the need to go back to the microstructure, and thereby to build up these sociological laws from the physical onesâ (Neurath 1931b/1983, p. 75). Cf. also Cartwright/Cat/Fleck/Uebel, 1996, Part 3: Unity on the earthly plane, pp. 167â252, esp. pp. 182â188.
âDie Materialistische Basis der Wissenschaftâ, Sozialistische Arbeitsgemeinschaft, 10 May 1930; âEinheitswissenschaft auf physischer Basisâ, Verein Ernst Mach, 20 May 1930; âDie Psychologie in Rahmen der Einheitswissenschaftâ, 28 May 1930. Psychologisches Institut Kolloquium.
Carnap 1931/1934 and Carnap 1932a/1959. Henceforth referred as âUniversalspracheâ and âPsychologieâ.
Carnap rc 110-03-22 asp and rc 110-03-36 asp.
An objection to such a translation may be that the indexical ânowâ has also to be translated into physicalistic language, but it is not clear how? I do not address this problem here, and since this is a quotation from Carnapâs text (see Unity of Science, p. 87.), it would be misleading to change it.
Cf. Carnap 1931/1934, p. 87.
Cf. Carnap 1931/1934, pp. 90â91.
It is worth noting that I do not intend to suggest that Carnap was an eliminativist or sceptic about experience (in agreement with Uebel 2018). I claim only that the physicalistic translation of psychological sentences need not appeal to a general psychophysical parallelism thesis.
Cf. Carnap 1932b/1987, p. 457.
Carnap 1932b/1987, p. 468.
It seems justified that such a broadly Humeian approach to persons could be credited to Carnap.
I argue for this in detail in Ambrus 2020.
A note on the term ânature of experienceâ and its characterization I ascribe to Carnap (and to Neurath in the next section). It may be objected that, in what follows, I am upsetting the common understanding and self-understanding of Neurath and Carnap by attributing to them substantial metaphysical views, i.e. âtheories of experienceâ or âtheories of consciousnessâ (even if only implicitly). Moreover, I even claim that certain linguistic theses, for example, Neurathâs claim that there cannot be a private phenomenal language is based on these âmetaphysical viewsâ. Hence I may be seen as interpreting Carnap and Neurath in a way that invites the objection that they are inconsistent in their avowed anti-metaphysical approach to philosophy, an objection similar to Armstrongâs, who accused linguistic philosophers of being closet-metaphysicians (Armstrongâs target was Ryleâs logical behaviourism. See Armstrong 1977/2002, p. 80). What I have in mind, though, is less radical, and, I believe, it can be accommodated with Neurathâs and Carnapâs metaphilosophical views, that is what they took to be the legitimate goals and methods of philosophy. I maintain that my characterization of their views about âthe nature of experienceâ are not constituents of a metaphysical account, but of an empirical one. Whether we have reliable direct/introspective capacities to diachronically re-identify our experiences, seems to be, partly at least, an empirical question. Whether the phenomenal character of our experiences is diachronically constant or not, also seems prima facie empirical. As such, admittedly, they belong to science, not to the proper subject-matter of philosophy, according to Carnapâs division. Nonetheless they may constrain in an indirect manner what can be formulated significantly (the investigation of which is a major objective of philosophy), hence it is legitimate to consider these issues in adjudicating what sort of languages are possible. To note: Ayer (1954), in his reply to Wittgensteinâs private language argument, seems committed to the assumption, at least implicitly that we have a reliable introspective capacity to diachronically re-identify experiences. See e.g. Ambrus 2021. In contrast, the issue whether experience is logically private may be seen primarily as a conceptual (or âgrammaticalâ) issue: cf. e.g. Wittgensteinâs arguments in the early thirties to the point the feeling another personâs pain is logically impossible, that is, a sentence with this content is âforbidden by the Syntaxâ (cf. Wittgenstein 1964, pp. 89â94). But see also Schlickâs contrary view (Schlick 1936, section v, pp. 358â369.) according to which the privacy of experience is only factual; it is not logically impossible to feel othersâ sensations. Accordingly, I intend to use the term in an innocent, empirical sense, without contradicting Carnapâs and Neurathâs meta-philosophical views.
Carnap 1932a/1959, pp. 177â179.
Cf. Waismann, 1967. pp. 49â50; Wittgenstein 1964/1975. pp. 88â94. See more detailed in Ambrus 2020.
It is interesting to note that Ayer (1956) accepted the logical privacy of experience but maintained both the possibility of private and public phenomenalistic languages. However, he interpreted the logical privacy of experience and the publicity of language differently from Carnap. According to Ayer, the logical privacy of experience is a linguistic convention that is based on the empirical fact that we have no direct access to other minds, while the criterion of the publicity of a language is not that its terms refer to public objects, but that the meaning of its words, i.e. the rules of their use, can be made intelligible to others, even if the language was created by a single person, âin solitudeâ; cf. Ayer 1956, Chap. v (iii) The privacy of experience; (iv) What can we communicate? pp. 226â238; see also Ambrus 2021.
Regarding this issue, I find myself in agreement with Uebel 2018, where he elaborates the claim that private phenomenalistic languages were considered possible by Carnap throughout his career, not only in his phenomenalist period, or in the beginning of his physicalist period, as I am investigating here; phenomenalistic languages are ineligible only as scientific languages. See later also.
E.g. in Uebel 1992, 1995, 2007.
Cf. Manninen 2002, 2003.
Cf. Manninen 2002, 2003.
âEinheitlichkeit der Gegenstände aller Wissenschaftenâ, Neurath undated/1930/1981.
Neurath 1931a/1983, p. 55.
Neurath 1931b/1983, p. 65.
Uebel 2007, pp. 228â229 (emphasis added â g.a.).
See Wittgenstein 1953, sec. 293.
Uebel 2007. p. 228.
I interpret the controller and the forecaster, in line with Neurathâs intended use, as two persons existing at different times who are psychologically and bodily connected and continuous, (i.e. that the constituents of their minds and their bodies overlap to some appropriate degree), that is, who would be regarded commonsensically as the same person at different times.
Emphasis added â g.a.
Someone may object that we have better access to our past selves than to other minds. The obvious reply from Neurath could have been been that this is a difference of degree not kind; access in both cases is indirect and unreliable, hence any such difference is therefore irrelevant. One may press the issue further, however, by arguing that there are decisive â phenomenological â differences between imagining or even âquasi-rememberingâ (in the sense of Parfit 1971) another personâs experience and remembering, recalling oneâs own earlier experiences (see e.g. Schechtman 1994, 1996, 2011 and Casey 1987). For, the argument goes, the content of memories are â in most, or perhaps all cases â personal, which is tantamount to saying that the contents of memory states are autonoetic: they include that the subject of the remembered experience is identical with the rememberer. But since there is no textual evidence that Neurath was concerned with such considerations, it is not to be expected that his account may provide answers to these possible objections.
Concerning the term ânature of experienceâ or ânature of consciousnessâ, see my remarks in footnote 17.
It may be asked whether experience is logically or empirically private, according to Neurath. In other words: is the proposition that one has no direct access to other minds, and in Neurathâs view, neither to oneâs own âpast selfâ, based on empirical or logical impossibility? Carnap, Hempel and Schlick discussed the problem of direct access to other minds in terms of logical possibility or impossibility, and I would venture this was the dominant approach among the logical empiricists. Whether Neurath also subscribed to this view, needs further elucidation. It is also possible that for him the issue was of no importance, or that he held the distinction could not be made precise in principle. In any case, the truth of claims (2) and (3) seems independent of this issue, and Neurathâs argument against private language could be formulated also if we assume that experience is only empirically but not logically private (or else, private simpliciter, if this distinction is not be applied in Neurathâs case).
Neurath would have rejected the terminology of âGivenâ â I am using it only to make the contrast more vivid between Neurath and Carnap, by formulating the views I ascribe to them by using the same terminology.
This phrasing does not mirror Neurathâs views in that it seems to assume that Neurath embraced a traditional version of the Verification Principle, what he arguably did not. Nonetheless, I hope, the content of his view may be correctly captured by (3).
One may object that this interpretation implies that Neurath was not an empiricist â which would, indeed, be a strange claim to make â because accepting (2) and (3) would imply that there is no possibility any longer to consistently and intersubjectively refer to the perceptible world. But I do not see why this would follow (2) allows that our experience, the phenomenal character of our perceptual states, may, in fact, have a diachronically constant nature (i.e. being the same or similar in like observational circumstances). But since we cannot verify this solely internally, such a claim has no truth value, hence cannot be true. In order for a person to be able to diachronically re-identify the content of her perceptions, she needs some external guarantee; hence perceptual contents must be interpreted as referring to public physical objects and properties. But to claim that the content of perceptual reports must be about observable physical objects and properties, not about the properties of experience is not to deny that only statements having some sort of empirical i.e. observational verification base are meaningful. Neither does it imply denying that all factual knowledge must be based on perceptual experience one way or another. So, it seems to me, my characterization of Neurathâs views by (2) and (3) is neutral on the issue of empiricism.
Cf. Carnap 1928/1967 p. vii, and Carnap 1963, p. 945.
Consider the following formulation of Carnap: âWhen I read the old formulations today, I find many a passage which I would now phrase differently or leave out altogether; but I still agree with the philosophical orientation which stands behind this book. This holds especially for the problems that are posed, and for the essential features of the method which was employed. The main problem concerns the possibility of the rational reconstruction of the concepts of all fields of knowledge on the basis of concept that refer to the immediately givenâ (Carnap 1961/1967 p. v). And âIt is an essential characteristic of phenomenal language that it is an absolutely private language which can only be used for soliloquy, but not for common communication between two persons. In contrast, the reistic and the physical languages are intersubjectiveâ (Carnap 1963, p. 869). These statements clearly show that Carnap upheld the view also in the '60s that phenomenalistic private languages are possible. It is, of course, puzzling first, how to reconcile this with his claim that the universal language of science ought to be the physicalistic language, he advocated since the âUnity of Scienceâ. However, as I suggested, this apparent contradiction can be explained away be suggesting that Carnap only proposed physicalistic language to be used for scientific purposes. This suggestion does not rule out that solipsistic phenomenalistic language is meaningful. Furthermore, according to Uebel 2018, Carnap, by 1935 at the latest, reaffirmed that phenomenal autopsychological sentences are verifiable directly via introspection or indirectly by behavioural indicators, hence they again, ought to be considered as scientifically meaningful again (cf. Carnap 1936/37). So my view is in line with Uebelâs detailed explanation concerning the different tasks what Carnap assigned to the solipsistic phenomenalistic language and the intersubjective physicalistic language. Cf. âIt is evident that Carnapâs abandonment of methodological solipsism remained intact [also in the late period when writing the introduction to the second edition of the Aufbau â g.a.)] [â¦] After all, Carnapâs abandonment only meant a discontinuation of the employment of phenomenalist languages for practical use or their rational reconstruction; they remained available as the objects of logical investigationâ (Uebel 2018, p 377; emphasis added â g.a.).
Of course, it is to be further investigated whether Carnap or Neurath is closer to truth on this point, further reasons may be discussed for and against the assumption that we have a reliable introspective capacity to re-identify our experiences, but I shall not pursue further this issue here.
Acknowledgements
I am grateful to Joseph Bentley, Christian Damböck, Johannes Friedl and Ulf Höfer for their valuable comments on an earlier draft of this paper, as well as to Jordi Cat, Hans-Joachim Dahms, Christoph Limbeck-Lilienau, Friedrich Stadler, Adam Tamas Tuboly and Thomas Uebel for their helpful and illuminating comments in the discussion of the presentation on which the paper is based, at the Wege der Wissenschaftlichen Weltauffassung. Rudolf Carnap und Otto Neurath conference in Graz, 26â27.09.2019. I would also like to thank for the support received in the frames of the K-120375 nkfi-otka research project of the National Research, Development and Innovation Office, Hungary and from the Higher Education Institutional Excellence Grant of the Ministry of Human Capacities entitled Autonomous Vehicles, Automation, Normativity: Logical and Ethical Issues at the Institute of Philosophy, elte Faculty of Humanities, in the period of working on this paper.
References
Ambrus, G. (2020) âCarnap and Wittgenstein on Psychological Sentences: 1928â1932. Some Further Aspects of the Priority-dispute over Physicalismâ, in Fisette, D., Fréchette, G. and Stadler, F. (eds.) Franz Brentano and Austrian Philosophy, Cham: Springer, pp. 353â386.
Ambrus, G. (2021) âThe Evolution of Ayerâs Views on the Mind-Body Relationâ, in Tuboly, A.T. (ed.), The Historical and Philosophical Significance of Ayerâs Language, Truth and Logic, Cham: Palgrave Macmillan/Springer, pp. 153â190.
Armstrong, D.M. (1977/2002) âThe Causal Theory of Mindâ, in Chalmers D. (ed.), Philosophy of Mind. Classical and Contemporary Readings, Oxford/New York: Oxford University Press, pp. 80â87.
Ayer, A.J. (1954) âCan There be a Private Language?â, Aristotelian Society Supplementary Volume 28 (1), pp. 63â76.
Ayer, A.J. (1956) The Problem of Knowledge, London: Macmillan.
Carnap, R. (1928/1967), Der Logische Aufbau der Welt ,Berlin: Weltkreis-Verlag, English translation by George, R.A.:The Logical Structure of the World, in The Logical Structure of the World. Pseudoproblems in Philosophy, Berkeley: University of California Press, 1967. pp. iâxx; 5â300.
Carnap, R. (1931/1934) âDie physikalische Sprache als Universalsprache der Wissenschaftâ, Erkenntnis 2, pp. 432â465, English translation: The Unity of Science, London: Kegan Paul 1934.
Carnap R. (1932a/1959) âPsychologie in physikalischer Spracheâ, Erkenntnis 3, 107â142, English translation: âPsychology in Physical Languageâ, in Ayer, A.J. (ed.) Logical Positivism. Glencoe, Ill.: The Free Press, 1959, pp. 165â198.
Carnap, R. (1932b/1987) âÃber Protokollsätzeâ, Erkenntnis 3, pp. 215â228, English translation by Creath, R. and Nollan, R. âOn Protocol Sentencesâ, Nous 21 (4), 1987, pp. 457â470.
Carnap, R. (1936/1937) âTestability and Meaningâ, Philosophy of Science 3, pp. 419â71; 4, pp. 1â40.
Carnap, R. (1963) âNelson Goodman on Der logische Aufbau der Weltâ, in Schilpp, P.A. (ed.), The Philosophy of Rudolf Carnap, La Salle, Ill.: Open Court, pp. 943â946.
Cartwright, N., Cat, J., Fleck, L. and Uebel, T. (1996) Otto Neurath: Philosophy between Science and Politics, Cambridge: Cambridge University Press.
Casey, E.C. (1987) Remembering. A Phenomenological Study, Bloomington/Indianapolis: Indiana University Press.
Feigl, H. (1963) âPhysicalism, Unity of Science and the Foundations of Psychologyâ, in Schilpp, P.A. (ed.), The Philosophy of Rudolf Carnap, La Salle, Ill.: Open Court, pp. 227â267.
Manninen, J. (2002) âWie entstand der Physikalismus?â, Nachrichten. Forschungstelle und Dokumentationszentrum für Ãsterreichische Philosophie 10, pp. 22â52.
Manninen, J. (2003) âTowards a Physicalistic Attitudeâ, in Stadler, F. (ed.), The Vienna Circle and Logical Empiricism. Re-Evaluation and Future Perspectives (Vienna Circle Institute Yearbook 10), Dordrecht: Kluwer, pp. 133â150.
Neurath, O. (undated/1930/1981) âEinheitlichkeit der Gegenstände aller Wissenschaftenâ, in Gesammelte philosophische und methodologische Schriften, vol. 1, edited by R. Haller and H. Rutte, Vienna: Hölder-Pichler-Tempsky, pp. 393â396.
Neurath, O. (undated/1930) âProto-Soziologieâ, Vienna Circle Archive, Haarlem, manuscript
K.2.
Neurath, O. (1931a/1983) âPhysikalismusâ, Scientia , Nov. 1931, pp. 297â303, English translation: âPhysicalismâ, in Neurath (1983), pp. 52â57.
Neurath, O. (1931b/1983) âSoziologie im Physikalismusâ, Erkenntnis 2, 1931, pp. 393â431, English translation: âSociology in the Framework of Physicalismâ, in Neurath (1983), pp. 58â90.
Neurath, O. (1932/1983) âProtokollsätzeâ, Erkenntnis 3, 204â214, English translation: âProtocol Statementsâ, in Neurath (1983), pp. 91â99.
Neurath, O. (1983) Philosophical Papers 1913â1946, edited by R.S. Cohen and M. Neurath, Dordrecht: Reidel.
Parfit, D. (1971) âPersonal Identityâ, Philosophical Review 80 (January), pp. 3â27.
Schechtman, M. (1994) âThe Truth about Memoryâ, Philosophical Psychology 7 (1), pp. 3â18.
Schechtman, M. (1996) The Constitution of Selves, Ithaca: Cornell University Press.
Schechtman, M. (2011) âMemory and Identityâ, Philosophical Studies 153 (1), pp. 65â79.
Schlick, M. (1936) âMeaning and Verificationâ The Philosophical Review 45 (July), pp. 339â369, available at https://doi.org/10.2307/2180487
Stoljar, D. (2017) âPhysicalismâ, in The Stanford Encyclopedia of Philosophy (Winter 2017 Edition), Edward N. Zalta (ed.), available at: https://plato.stanford.edu/archives/win2017/entries/physicalism/.
Uebel, T. (1992) Overcoming Logical Positivism from Within ,Amsterdam/Atlanta: Rodopi.
Uebel, T. (1995) âPhysicalism in Wittgenstein and the Vienna Circleâ, in Gavroglu, K. et. al. (eds.), Physics, Philosophy, and the Scientific Community, Dordrecht: Kluwer.
Uebel, T. (2007) Empiricism at the Crossroads: The Vienna Circleâs Protocol-Sentence Debate, Chicago: Open Court.
Uebel, T. (2018) âCarnapâs Transformation of Epistemology and the Development of His Metaphilosophyâ, The Monist 101, pp. 367â387.
Waismann, F. (1967) Ludwig Wittgenstein and the Vienna Circle. Conversations Recorded, edited by B. McGuinness, transl. J. Schulte and B. McGuinness, Oxford: Blackwell.
Wittgenstein, L. (1953) Philosophical Investigations, edited by G.E.M. Anscombe and R. Rhees, transl. G.E.M. Anscombe, Oxford: Basil Blackwell.
Wittgenstein, L. (1964/1975) Philosophical Remarks, transl. R. Hargraves and R. White, Blackwell: Oxford.