Press "Enter" to skip to content

Has private language theory been revived?

A reassessment of Wittgenstein’s arguments against private language, with respect to the current cognitive climate, in an attempt to ascertain what (if any) comparisons can be made, and to what extent a relationship can be made.

© 2001, Alistair Knock. This essay may be redistributed in its original, unmodified form, in whole or in part, providing due recognition is given to the author.

Ludwig Wittgenstein’s exploration and denouncement of private language is perhaps the most interesting and accessible part of the Philosophical Investigations, largely because it is one of the few aspects of philosophy that Wittgenstein provides a clear response to: that private languages are impossible. By private language, he is referring to the possibility of one utilising a language that no other individual could or should comprehend, and thus that the language lacks a public perspective. The principle weapon that Wittgenstein carries is precisely the latter — that language by definition finds its purpose in being communicated from one individual to another, and so instantly denying the possibility for a language that is confined to the self. However, the philosophy of language has seen itself travel some distance on the wave of cognitive science, and much of the accepted philosophy of today bears little resemblance to how Wittgenstein’s post-Vienna world looked. It is upon this distinction that this essay is founded: to attempt to reconceive the arguments made by Wittgenstein about private language, in the context of what a modern-day version of Wittgenstein might resemble.

We shall firstly analyse the text Wittgenstein presents us with in the Philosophical Investigations, (though many commentators isolate sections 243-275, the argument meanders and is integrated into Wittgenstein’s entire discussion) and extrapolate his own definition of a private language, and how he rejects the possibility of it. Having established a more concrete picture of how Wittgenstein sees the private language battlefield, we shall introduce some popular contemporary (and some traditional) philosophies, and engage them with the private language encountered with Wittgenstein. In doing this, there are some key questions we require an answer to (each from the perspective of both Wittgenstein, and our alternative philosophers): What is language? What is private language? What is our concept of thought? How significant is logic in language, and thought?

Much of Wittgenstein’s later work is dedicated to what he calls the ‘language-games’ that we play, which emerge primarily from the imprecision of languages as we know them. We are all aware of the result of one’s inability to express their thoughts adequately through language — the time-honoured response, ‘what do you mean?’ This, to Wittgenstein, is not to imply that the first person’s meaning is lacking, but rather that the expression of it does not provide enough information to re-formulate the thought in another’s mind. He does not explicitly break down language into grammar and words, but speculates that having known the way one language works, one could by all means learn or pick up an entirely unrelated language — this hints that to Wittgenstein, there is a common grammar in languages, at least in the way sentences and phrases are constructed. This is of interest to us when we later consider the work of one contemporary proponent of cognitive science, Jerry A. Fodor. The crucial feature of language that Wittgenstein emphasises though, is that it is necessarily shared — that without assertions and denials and justifications between more than one party, the purpose of language is lost. ‘Language is, therefore, essentially a communal language.’

Though Wittgenstein details a somewhat pictorial concept of thought throughout the Tractatus, he does not deal with concrete concepts in the Investigations — this is likely to be because he now involves himself in the objectivity of the issue, and does not wish to postulate concepts which cannot, by definition, be discovered. Broadly, however, we may take a generalised position: that thought is the workings of the mind. This presents us with an interesting question that is not touched on in his writing — whether he permits thought a sub-conscious layer, or whether it is purely conscious and ‘known’ to our alert and awakened self. The inclusion of the sub-conscious in our discussion is integral, since most of the alternative models we shall look at involve, or at least recognise, the sub-conscious in some way. Wittgenstein’s stance differs here — although he is familiar with Freud’s work in the division of the mind, it does not show in the Investigations. Whenever the mind or thought is referred to, it is done so without any further investigation, as though Wittgenstein either finds it unnecessary, or overlooks, the possibility of what essentially amounts to multiple minds. By this I mean the theoretical division and communication between parts of the mind, as opposed to any physical, scientific establishment — Freud’s division of the mind is for analytical and treatment purposes of other behaviour rather than as an explanation of mind activity, but nevertheless the a priori possibility remains that a mind could be split into rival or co-operative factions. Wittgenstein does provide us with an insightful précis of his idea of the interrelation between language and thought: ‘When I think of language, there aren’t ‘meanings’ going through my mind in addition to the verbal expressions: the language is itself the vehicle of thought.’ This seems to entail that both language and thought are one, an intertwined dualism. As Daniel Laurier expresses in his essay entitled The Publicity of Thought and Language, it is common for us to think of language extending from thought, but the reverse portrayal, that thought is produced as a result of language, is somewhat more difficult to grasp. The direct upshot of this engaging statement by Wittgenstein is the assumption that there is no thought that is not founded on language, and thus that there is no thought that cannot be communicated.

In sections 243 to 275, he introduces his definition of a private language: ‘the individual words of this language are to refer to what can only be known to the person speaking; to his immediate private sensations. So another person cannot understand the language.’ We should take care to notice that the words refer to what can only be known to the self, i.e. that it is the meaning, the semantic, that is private — the words used are not some sort of mythical internal mode of communication, but rather representative tokens of the private meaning. Thus it could immediately be argued that, in some circumstances, a private language could indeed be communicated – it would just lack any meaning or purpose for the recipient, nullifying its use as a language. Wittgenstein focuses his discussion of private language on the language of sensations — pain, happiness, and so on — whereas much of the Investigations uses everyday public objects as its foundation. This is interesting because it hints that Wittgenstein seems to feel the realm of private language is in what is immediately recognisable as private, such as emotion, but the fact Wittgenstein singles out sensations in his dealings with private language perhaps draws too small a picture — also, the area of mind activity he focuses on primarily, sensations, is a passive feature of the mind, without entering into the opposite side, that of action. Whether these limitations damage the concept of private language that Wittgenstein revokes remains to be seen. To clarify, ‘what Wittgenstein had in mind is a language conceived as necessarily comprehensible only to its single originator because the things which define its vocabulary are necessarily inaccessible to others.’

After hopefully conceiving his definition of private language, Wittgenstein revokes it as being, in effect, impossible. This rejection is concentrated in section 258, and is summarised well by Stewart Candlish: his ‘conclusion is that a language in principle unintelligible to anyone but its originating user is impossible. The reason for this is that such a so-called language would, necessarily, be unintelligible to its supposed originator too, for he would be unable to establish meanings for its putative signs.’ Wittgenstein is discussing the case whereby when one experiences a certain sensation, he makes the mark “S” in a calendar for every day the sensation is felt. A problem emerges, in that this symbol does not give a definition of the sensation, nor does it provide any meaning or explanation of the sensation. Since the purpose of this exercise is to establish whether the presence of a ‘private language’ necessarily enables normal activity of the mind, Wittgenstein discovers a loophole, as it were — for on every subsequent occasion that this sensation is experienced, there can be doubt over its authenticity as the same sensation, since the use of private language has not provided us with an adequate meaning or definition of the sensation to reference back to. Thinking about the connection between sign “S” and the sensation as it occurs means only that ‘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 can’t talk about ‘right’.’ Here we should note that Wittgenstein does make a definite distinction between the representative symbol and the meaning behind it, but seems to imply that with a private language one could not exist without the other, and that the connection linking the two is crucial. If such a language was truly private, the connection would be lost. He does not appear to delve deeper into this cognitive circumstance, which is unfortunate, since there are alternative situations whereby the meaning is purely private and subjective, yet the connection and symbol are public, or where all three are publicly available but interpreted in a subjective, private way, and so on. These omissions we shall focus on more clearly in later discussion with other philosophers.

The claim here is that private language as Wittgenstein sees it, will only work if the originator of the language is fully exposed to the symbols and meanings used in the language — and that, quintessentially, if this is the case, there is little to prevent this private language becoming public, through communication to other individuals. ‘What Wittgenstein aimed to show is not that sensations language, like the rest of language, is essentially shared, but that it is essentially shareable.’ Wittgenstein of course is influenced by what we have already discovered with respect to his opinions on language, that it is by its very nature and operation, communal, and so it is hardly surprising that he wishes to reject a language that intends to keep to itself within one individual. It is from this setting, that we proceed.

The cognitive revolution admittedly took firm hold after Wittgenstein’s time, but that is not to say that he was excluded from its early days. Sigmund Freud, in the early 20th century, conceived of his famous tri-partite view of the mind into id, ego, and superego. Indeed, this formulation has its roots in Plato’s division of the individual, as seen in Republic. Though these texts did not deal directly with our subject, nor to the depth that we require them to, they are significant because they show that the idea of multiple minds per individual was not a post-Wittgensteinian development. By multiple minds, I refer to the possibility of the mind or soul consisting of several autonomous, sentient, constituent components. Such components would possess the ability to consider and ‘think’ independently of the others, and notably could conflict or agree, by means of communicationinternal to the mind. This type of model is the route that cognitive science has embraced in the late 20th and early 21st century, and poses a very interesting alternative perspective on Wittgensteinian private language — for a language self-contained within the mind but shared between several participants is communal, and yet could still remain obscured from the view of the sentient, conscious, ‘real mind’ (by means of the conscious/sub-conscious divide). The prevalence and general acceptance of contemporary philosophies that ground themselves in some sort of language of thought seems a sharp thorn in the side of those that deny private language, and the success or weakness of such opposing theories is our primary focus.

In the 1970s, there emerged a philosophical shift toward representational theory that is largely the result of developments in the field of computing, the computational model, and in part a backlash against the overwhelming success of behaviourism. There are of course differing factions, such as those pursuing ‘intentional realism’, or the Language of Thought hypothesis (LoT) conceived of and advocated by Jerry A. Fodor, but essentially the underlying foundation of a representational theory of thinking is this: when dealing with thoughts, beliefs, desires, and so on, the mind deals with representations of the objects or external states themselves, and these are physical representations that possess semantic content. Essentially, when we think of raising our left arm, our mind locates the representative tokens that stand for ‘raise’, ‘left’ and ‘arm’, analyses the semantic content stored within them, and acts appropriately according to the result of this complex combination. Here we should note that to representationalists, thoughts are complex sequences constructed from simple objects or tokens, rather than the opposing view that each individual activity or intention is distinct and unique (think, for example, of the activity of raising one’s left arm and hopping on one’s right leg simultaneously). As Fodor notes in his groundbreaking book, The Language Of Thought, the introduction in the latter half of the 20th century of functioning computers capable of performing complex tasks and, crucially, equipped with the possibility of exhibiting some sort of intelligence, encourages philosophy to refocus itself around these developments. Computers traditionally operate on two levels: the first, approachable level is the human-computer interface — the means by which one communicates with the computer itself. This is primarily through such devices as mouse and keyboard, but firstly a program is required, and this is our main concern. All computers are programmed in a language that both human and computer understand — the mode of communication is thus a public one. However, once received by the computer, it is translated or interpreted into the computer’s ‘native’ language, i.e. the method by which it processes and communicates within itself. This native language is by definition, the way a computer ‘thinks’. Three crucial points must be noted before we consider this further (largely these are the result of Fodor’s response to attacks on his thesis):

  1. A computer is designed for the purpose of this native language. It is not designed to have the capacity to learn the language, it is designed to be the language.
  2. It is neither necessary for the computer to be aware of the language it uses to communicate and process within itself — the familiar input-process-output model advocated by John von Neumann has spawned the concept of the black box as process, whereby the central activity is unimportant so long as the input and output are gained.
  3. There is conflicting opinion as to whether this is a comparable instance, primarily over the question of whether, like machines, humans are designed. This is easily overcome: for the traditionally religious, God has the power to implement private language; for staunch Darwinians, Paul and Patricia Churchland venture into interesting avenues of the luck of evolution. . Though they conclude that a language of thought is likely to be an outside bet, much of this conclusion is based on rationality, and avoids the stark fact that we have evolved from amoebas into conscious, sentient, geniuses. That there is luck involved in this equation cannot and should not be denied, but luck and probability does not single-handedly defy the possibility of representationalism or language of thought.

The reason why a computer requires an internal language is because it consists of individual components, each of which operate at times independently of one another, sometimes in co-operation with other parts. A common method of communication is thus a necessity — but it should be noted that the computer is dysfunctional without the co-operation of all its constituent parts. If one element is removed, the powers of the machine are lost, and so we should still depict the concept of computer as a single entity, which is dissolvable into its fundamental parts. To most contemporary cognitive scientists, philosophers, and psychologists, the same is true for the mind. We have already touched on the concept of multiple minds, which comes in a variety of flavours. Daniel Dennett introduces the somewhat idealistic notion of multiple draft theory in his volume, Consciousness Explained. Here the mind is fragmented into independent systems which deal with particular aspects of ordinary human existence — retrieval of memory items, check-ups on the status of the hand, and so on — but there is no overlord, no decision maker to judge or arbitrate the priority or acceptance of particular results. Instead, each system deposits its output into a central pool of ‘drafts’, which are then added to by other systems or updated by the originator: the success of this output is dependent on how long the draft is sustained by contributing systems — if the originator stops supporting the output shortly after its creation, it will fade, and the mind will focus on another candidate. The detail of how such an operation can be successful lacks plausibility, but nonetheless we can see how an internal common language for these systems is again necessary. (Note that the Cartesian model of the mind bears some similarities to this conceptualisation, with the major difference being the inclusion of a ruling system: the soul, accessed through pineal gland. Descartes’ greatest obstacle is the problem of communication between ethereal soul and material body.)

Similar work has been pursued my work on mindware, which accepts the notion of constituent systems of the mind, but avoids the hit-and-miss symptoms of Dennett’s proposal, as well as evading the inclusion of a ruling body. Instead, mindware theory proposes that constituent systems evolve and interact according to something akin to a rulebook or guiding framework. Their method of communication and of operation stem from one, passive source (one trouble-free way to visualise this is to think of DNA) which plays no formal role in the operation of the mind, but instead shows constituent systems how best to operate. The primary elements of mindware are prioritisation, organisation, and communication, with the desired result being increased efficiency and effectiveness. That internal language plays a crucial role between these multiple minds is beyond doubt. (Note that like DNA, the constitution of mindware varies widely from person to person, thus there are no two minds that possess the same construction)

We should answer the brief we set ourselves at the beginning of this mission. For the contemporary philosopher, (conscious) language is possibly much like Wittgenstein describes it: the vehicle of thought. There is an irony in this statement, for secondly, many colleagues will maintain that it is also literal — that physically, the method by which the mind deals with thought, is language. Whether this internal language is a private language in the sense that Wittgenstein perceives it, we shall investigate, but Fodor certainly expresses enthusiasm over the prospect, in naming a section of The Language Of Thought, ‘Why There Has To Be A Private Language’. With the definition of thought, contemporary philosophers are certainly affected by computational theory. Though there are many who continue to abide by folk psychology, developments both electronic and in neurotics have assisted the cause of cognitive scientists looking for a physicalist array of constituent, co-operative systems. In the main, however, we can certainly stipulate that thought is not merely an activity confined to the conscious, although Wittgenstein does not himself touch on the subject, and that there are indeed several types of thought which correlate to brain activity. Quite whether we should involve any non-conscious activity as part of our comparison is a matter of personal opinion — personally, I believe the unearthing of the sub-conscious stratum invokes a wide-ranging threat to the majority of pre-20th century philosophy. The final two questions we postulated, on logic and the distinction between syntax and semantic, we shall deal with more fully.

Whatever the particulars of different modalities, the current philosophical picture of the mind does appear to differ from Wittgenstein’s formulation. This is difficult to substantiate because of his apparent lack of interest in subliminal brain activity contra Freud, but nevertheless Wittgenstein makes no effort to involve himself in the dynamics of a multiple mind system, and so we must assume his ignorance/rejection of such a possibility. Our task now is to ascertain whether private languages as perceived by Wittgenstein retain their integrity within the framework of recent cognitive developments. Let us recount the major points made by each side:

  1. Public language is necessarily communal and by its nature, shared.
  2. Private language is internal and refers only to what can be known to the self; thus another cannot understand a language: it is necessarily non-communicable. (This is not to say that the entire language is non-communicable, since Wittgenstein admits the feasibility of accessing the syntactical vocabulary, but semantics and meaning are essentially private).
  3. Private language is impossible, since a language unintelligible to external parties would be unintelligible to the originator also, because the analysis of such a language would lack the non-transferrable meanings of its vocabulary.
  4. If such a language was accessible by the originator, then both words and meanings must be approachable; and so a third-party could possess the ability to access them too, defying the privacy of the language.
  1. With respect to thought activity and intentionality, the mind deals purely with representations of objects, both internal (i.e. imagination) and external (based on sense datum).
  2. Simple objects are sequenced together to form complex objects; this is done to various degrees within the representational school of thought, with the extreme being the Language of Thought hypothesis which depicts brain activity as having syntactic sentence-like structure.
  3. Such a language could be thought to be private, since it deals with what the originator knows and has access to semantically. This is partly a matter of subjectivity, but also that the operation of the mind entails use of this language — without the unique mind itself, the language is incomprehensible.
  4. Thus such a language is private to the originating mind, since it has total dependence on the originator.
  5. It is not necessary for the mind to interpret and understand the language that it is analysing: see analogy with computer systems; the mind is the language, and so there is no need to regress further.

For private language to work for Wittgenstein, he demands it be inaccessible and cryptic for prying observers, yet it must be fully comprehensible to the originator, otherwise its purpose is lost. He feels that private language’s failure then, is the unattainability of such a ‘locked-away’ language. The subjectivity of the issue, that everyone has different meaning for the same objects, is a secondary attack, separate from the first: Wittgenstein does not see the need to go this deep; the point is won from a layman’s perception of the mind. However, Wittgenstein’s denial of this particular private language rests on the assumption that the originator has cause to observe the language that it uses, which incites the theory of infinite regress — an originator of a language observing the language it is originating, must have some method of thought that does not involve the observed language, and so on ad infinitum. The problem for Wittgenstein is that if a private language was not private, and could be observed by third parties, then the language is accessible and cohesive — but is obviously no longer private. When kept private, third parties cannot be involved in the observation, and the originator using the language cannot observe itself.

The solution to this particular problem is to remove the necessity for observation. As we have seen in our discussion of Fodor and comparison with computer systems, a computer that is built on the basis of its internal language has no need to interpret such a language further, or even acknowledge its existence. Whatever the design of our brain and mind circuitry, we as humans operate and function successfully without a working knowledge of ourselves — if we do have a language of thought, we are certainly not consciously aware of it, and have no need to be. Wittgenstein’s primary denial of the possibility of private language rests on the necessary incomprehensibility of such a language to the originator, without fully exploring just how necessary the comprehension of a language of thought need be. His fixation with keeping everything above board, as it were, on a conscious level, proves very restrictive in our considerations. Though Wittgenstein denies that private language is conceivable, he says little about what should replace it — whether the language discussed is in fact the way he sees the operation of the mind, just without a privacy clause, or whether he believes the mind is constructed on very different premises. Again, the relevance of these conceptions is paramount to understanding what Wittgenstein means, and in what context — with such a blurred environment in which to work, it is difficult to accurately attempt to integrate contemporary views of private language into Wittgenstein’s framework.

Should Wittgenstein have come across and fully considered the concept of multiple minds within an individual, with an attached language or other characteristic communicative mode as we see in the representational stance, it seems likely that he would still revoke the possibility of privacy. Through consideration, we immediately see that the main problem is that when multiple minds communicate, they understand and gain something from what is being said. This fits with Wittgenstein’s formulation of public language, but not with private, since the language is shared. Another individual will no doubt find it impossible to understand and communicate in the language of another mind, and so on a human, large scale level, this language is indeed private. But to each component within the mind community, the language is necessarily shared and public. Where should we stop? That, it would appear, is the crucial consideration. Wittgenstein’s work is admirable and philosophically evocative, but outdated even in its own time because of a disregard toward alternative possibilities. Both Plato and Freud have themes of argument in their formulation of the mind — indeed, inner conflict is the essence of human nature, between rationality and emotion: we must ask ourselves, did Wittgenstein pick this up, and if so, why has he chosen to avoid the issue?

It is no secret that the adoption of a second language is far less painful and swifter than when learning to talk for the first time, and many linguists have picked up on this fact, hypothesising that all natural languages are based around a root architecture, grammar, which changes little from tongue to tongue. Though the complexities of gendering in French (such as le and la) are not replicated in English, for instance, such differences preclude the real grammar, which I propose is logic. Combinatorial statements, such as ‘apple and pear’ presuppose an understanding of the logical operative of ‘AND’, and an analysis of any natural language will show involvement of predicate and subject. In short, language appears to ground itself on logic: logic is the mother tongue. The relevance of this to our discussion becomes apparent with the distinction of grammar and words — anyone can speak another language with a knowledge of the words used as representatives for nouns, verbs, and so on, so long as they also understand the grammar (which we can now presuppose for anyone utilising language). This means that if an individual is in a position to observe the internal language of another person, over time they may come to learn the language itself. However, this does not imply that the observer has a full understanding of the language, for as we have touched on, the subjectivity of internal thought is a bulky obstacle to any understanding of another mind — it is no longer a matter of pointing to an apple and saying ‘apple’ or ‘le pomme’, because from this moment on in an individual’s lifespan, apple gathers a vast amount of semantic meaning which is unique to that individual. On this basis, the private language argument in many ways is over before it is begun, if we immerse ourselves fully in the semanticity of the issue — the language of the mind cannot be anything other than private in terms of meaning, since the language and the thoughts of each individual are co-dependent and inseparable.

There is undoubtedly a visible difference between the overall philosophical perceptions of mind affairs on each side, and for this reason it is perhaps unfair to subject Wittgenstein’s arguments on private language to thinking much further along the timeline of cognitive science. However, it should be noted that the arguments represented within this text that support the possibility of LoT or private language do not ground themselves in any way on a superior knowledge of the workings of the mind, either from a philosophical perspective or by way of developments in neuroscience. These advances certainly ease the explanation and acceptance of such hypotheses, but are not the foundations of them: ultimately, our discussion has been of an a priori nature. What we find lacking in Wittgenstein may be the result of the unfortunate and untimely death of a great philosopher who was unable to fully develop and explain his thoughts on the matter; or as has been noted in the past, perhaps there isn’t a private language argument in the Philosophical Investigations as such that we can analyse. The spectre of misinterpretation of course looms above us constantly; however, despite the efforts of Wittgenstein and his proponents, the denial of private language does not appear to have stuck in contemporary philosophy — it is either dismissed or avoided altogether as an unimportant technicality. Ironically, it may be the very fact that we cannot introspect on our own usage of an internal and private language of thought, combined with the notion that it is subjectivity and meaning that permit privacy, which certifies the eventual collapse of Wittgenstein’s attempt to quash private language. It is Wittgenstein himself who reasserts the subjective stronghold of the mind: ‘I act with complete certainty. But this certainty is my own.’


On Wittgenstein —

  • Stewart Candlish, Stanford Encyclopaedia of Philosophy: Private Language, available at (1998)
  • Robert J. Fogelin, Wittgenstein: Second Edition, Routledge (1976)
  • Oswald Hanfling, Wittgenstein’s Later Philosophy, MacMillan Press (1989)
  • Saul A. Kripke, Wittgenstein On Rules and Private Language, Blackwell Publishers (1982)
  • Daniel Laurier, The Publicity of Thought and Language, submitted to 20th World Congress of Philosophy, available at
  • Marie McGinn, Wittgenstein and the Philosophical Investigations, Routledge (1997)
  • Ludwig Wittgenstein, On Certainty, Blackwell (1969)
  • Ludwig Wittgenstein, Philosophical Investigations, Blackwell Publishers (1953)
  • Ludwig Wittgenstein, Tractatus Logico-Philosophicus, Routledge (1922)

On representationalism —

  • Jerry A. Fodor, The Language Of Thought, Harvester Press (1976)
  • Daniel C. Dennett, Consciousness Explained, Allen Clement (1992)
  • Alistair Knock, Mindware: Why Marie is Marie, unpublished (2001).
  • Alistair Knock, Present and evaluate the arguments Jackson develops in ‘Epiphenomenal Qualia’, unpublished (2000).
  • Steven Pinker, How The Mind Works, Penguin Books (1997)
  • Anthologised in Mind And Cognition, ed. William G. Lycan, Blackwell (1999):
    • William Bechtel, The Case For Connectionism
    • Paul M. and Patricia Smith Churchland, Stalking The Wild Epistemic Engine
    • Daniel C. Dennett, True Believers: The Intentional Strategy and Why it Works
    • Jerry A. Fodor, Why There Still Has To Be A Language Of Thought
    • Jerry A. Fodor, A Theory Of Content
    • Tim Van Gelder, What Might Cognition Be, If Not Computation?
    • John Heil, Privileged Access
    • U. T. Place, Is Consciousness A Brain Process?

Discuss this article on’s forums