Introduction

The title of this article is an allusion to Karl Popper’s The Two Fundamental Problems of the Theory of Knowledge ([1979]2009). Popper considered them to be the problem of induction and the problem of demarcation. We shall discuss two other problems that are, arguably, even more fundamental, namely the problem of negation and the problem of the grounds of sensory experience. These have been resolved by Ludwig Wittgenstein.

It has been argued that Popper’s two problems belong to the field of philosophy of science and concern the justification of knowledge, while Wittgenstein’s problems belong to the field of philosophy of mind and concern the question of what the mind is. For at least two reasons this subdivision is misleading. First, one can use Wittgenstein’s ideas to criticize Popper’s solution of the problem of demarcation, for Wittgenstein offered a sound alternative to Popper’s discussion of how empirical propositions are distinguishable from metaphysical ones. Secondly, Popper’s criticism of empiricism and his (critical) rationalist alternative are partly grounded in a philosophy of mind that he inherited from representative idealism (of which both empiricism and rationalism are examples). Wittgenstein repudiated representative idealism for good reasons, implying that his insights gained within the field of philosophy of mind have serious consequences for philosophy of science. Since we have elsewhere (Smit 2015) discussed how Wittgenstein’s insights can be used to criticize Popper’s ideas, we will focus in this article on the question of what the implications are of Wittgenstein’s resolutions for how we should conceive of phenomena in life science.

The ideas we will discuss and their relevance for life science are radically different from current ideas of biologists and philosophers of biology. Unfortunately, they are sometimes difficult to grasp. To understand Wittgenstein’s resolutions we need to abandon dominant ways of thinking, and it has been remarked that that is as difficult as giving up an addiction.Footnote 1 We hope to show that, by discussing the resolutions of the two fundamental problems, it is worth the effort.

The Intentionality of Thought

Wittgenstein’s discussion of the problem of negation is revolutionary because he linked it to the intentionality of thought. Before we explain this link, we will first mention some conceptual features of the intentionality of thought by discussing some examples.

Humans can think about items or events not currently present to their senses. I can think of my daughter Judith studying frogs in Gamboa (Panama), of how Willem of Orange was in 1584 shot by Balthazar Gerards in Delft (the Netherlands), or wonder whether the dikes in the Netherlands have to be raised since sea levels are increasingly rising because of global warming. I can even think of something that does not exist, like a ghost or Father Christmas. Examples like these show that, in thought, we can represent things and events in the present, past, or future.

Another feature of the intentionality of thought is that we can truly think that things are thus-and-so, but also falsely that they are thus-and-so. A similar observation can be made with regard to expectations, desires, hopes, and intentions. I can expect A to come before she comes (and A may not come at all), desire a certain food item (although the item may be sold out), hope it will rain tomorrow (although it may be sunny), or intend to watch a movie in the cinema (although the movie may no longer be shown). These examples show that thoughts, expectations, and so on, can anticipate something even when what we think, expect, and so on, may not exist or occur.

In our times, a discussion of the intentionality of the mental is mainly associated with Brentano ([1874]1995). Yet ideas about intentionality can be traced back to Aristotle and the scholastics (see Kenny 1993; Lagerlund 2007; Hacker 2013a, Chap. 2). Aristotle and the scholastics distinguished intentional from natural existence. If we think of a horse, what it is for something to be a horse (the form) exists in our mind (intentional existence)Footnote 2 and may or may not exist in nature (natural existence). Brentano made a similar distinction and argued, in opposition to Descartes who argued that consciousness is the mark of the mind, that the intentionality of thought is the defining feature of the mental. Mental phenomena are, according to him, always directed upon an object. Yet what is meant by the phrase “directed upon an object”? Suppose that I believe or hope that Mary will come, then “that Mary will come” (called by grammarians the nominalization-accusative) is what I believe or hope no matter whether the belief is true or the hope fulfilled (the intentional content). Notice that the “intentional content” is here not literally an “object.” When Brentano speaks about the object of thought, the mental existence of an object does not refer to a thing. An object in the outside world is something to which we can point and can grasp and handle, unlike an object in the mental world. When I hope that Mary will come, then “that Mary will come” is an answer to the question what I hope: it is a grammatical object.

Brentano argued that all mental phenomena are directed upon an object. However, by discussing two examples, we will show that intentionality is not a feature of all of them. Feeling a pain is the first one. Brentano argued that when we have a sensation like a pain, we feel something, and that implies according to him that our sensation has an object, just as believing or hoping has an object. But although “pain” in the phrase “to feel pain” is a grammatical object specifying what I feel, that does not make “feeling a pain” intentional, for we can only feel a pain if we have a pain, while we can believe or hope something that may not exist or occur. We must not conflate here the grammar of a transitive verb (feel) with the grammar of an intentional verb (believe, hope, and so on). Of course, when we have, for example, a toothache, we can attend to our pain, but an attentional object is not an intentional one. Phantom pain does not show that feeling a pain has intentional content either, for we do not merely think that we have a pain, and the pain seems to be where our amputated leg would have been.

A second example showing that intentionality is not the mark of all mental phenomena concerns cognitive verbs such as “know” or “be conscious of.” Although these cognitive verbs are “directed towards an object,” they do not involve the possibility of mere intentional existence. Cognitive verbs like “know,” “be conscious of,” or “be aware of,” are factive: their objects exist or obtain “in reality,” not merely “in thought.” There must be something that we know or are aware of. By contrast: when I believe something, that may involve mere intentional existence, for my belief may be false.

These two examples show that, in contrast to what Brentano argued, intentionality is not the essential mark of the mental, for it is not a feature of all mental phenomena. The lesson we can draw is that we have to differentiate between mental phenomena since the grammar or logic of the various mental concepts is different.

The Problem of Negation

When I think of Judith or Willem of Orange, then my thoughts seem to point to someone who is in Gamboa or was in Delft, and when I think of Judith observing frogs in Gamboa or of Willem of Orange being shot by Balthazar Gerards in Delft, then I think of two states of affairs to which my thoughts seem to reach out. Hence, it seems that, in general, in thought I connect signs (or names) with persons (or objects, etc.) and project propositional signs (when I think the proposition that p, p is called a propositional sign) on states of affairs (by thinking that p, p represents the actual state of affairs I mean). We can ask, as Wittgenstein did in his early work, how a proposition can be false but meaningful. He called it the problem of negation (Wittgenstein 1961, p. 30): how can a proposition depict or represent what is not the case? This problem was already discussed in ancient philosophy. For instance, Plato, in his Theaetetus (189a), asked whether there is anything for us to think when we think falsely. For what we think then is not the case, so it appears that there is nothing here for us to think. The problem is, of course, that we can think what is not the case. How is this possible?

In the Tractatus Wittgenstein ([1922]1973)Footnote 3 argued that when we think that p, we think a possibility that may or may not be actualized. This possibility is specified or individuated by what we believe, namely that things are so.Footnote 4 When we believe truly that things are so, then what we believe is the case. It does not fall short of it. Yet this raises the problem of negation: suppose that we truly assert that not-p, how can it be that this proposition represents how things are in reality, that is, that it is not the case that p? Is there anything for us to think (as Plato asked in the Theaetetus)? It is clear that p does not obtain, but it is also clear that the proposition not-p does not depict or represent something other than p, for if it did, then “not-not-p” would not have the same sense as “p”.Footnote 5 If, for example, “This tulip is yellow” is false since the tulip is purple, the negative proposition (“This tulip is not yellow”) does not signify that the tulip is purple. Thus, the positive proposition seemingly foreshadows the fact that would make it true were that fact to obtain, and the negative proposition is parasitic on the sense of the positive one, and, as it were, reverses its sense. The very same fact that verifies p also falsifies not-p. Wittgenstein argued that there is an essential connection between the two truth-values of an empirical proposition. And he argued that negation signifies an operation, not a function, implying that negation introduces nothing new into the proposition. The point to notice here is that Wittgenstein’s ideas are radically different from those of his predecessors Frege and Russell. We shall briefly spell this out, for it will aid us in understanding the differences between empirical, logical, mathematical, and conceptual propositions.

Frege, as is well-known, revolutionized logic. He developed a function-theoretic logic as an alternative to traditional syllogistic logic. The essential innovation was that the concepts of subject and predicate were in Frege’s new logic displaced by argument and function. Frege’s innovation was based on the generalization of the mathematical theory of functions: arguments and values of functions were no longer solely applied to mathematical objects, but were extended to all kinds of objects. This enabled Frege to argue that, from a logical point of view, what is denoted by, for example, the expression “The capital of φ” is no less a function than x2 + 2. In the latter the function takes the value 6 for the argument 2; in the sentence it takes the value Amsterdam for the argument the Netherlands. The sentence as a whole is, in Frege’s logical notation, to be viewed as the name of a truth-value. The truth-value of a sentence is the circumstance that it is true or false, which Frege called “the True” and “the False” for brevity.

Simple sentences are in Frege’s system functions from objects to truth-values. But Frege went further and construed in his system the logical connectives (negation, conjunction, disjunction, and conditionality) as unary or binary functions. Consequently, they became functions from truth-values to truth-values, for the logical connectives are in Frege’s system themselves (first-level) functions mapping objects onto objects. Wittgenstein criticized Frege’s function-theoretic account for many reasons. We discuss three of them (other reasons are discussed in Baker and Hacker 1984; Baker 1988; and Kanterian 2012).

First, in Frege’s logic, the proposition is not essentially connected to both truth-values. If, for example, F(a) expresses a true thought, its sense is that the True is the value of the function F(x) for the argument a. But what is the relation to the False? All that can be said is F(a) designates something different from the False (for it is not the name of the False). But nothing clarifies in Frege’s system that the operation negation reverses the sense of the positive proposition.

Second, if “not” were a function-name, then “not-not p” would say something different from “p”, for “not-not p” would be about not and “p” would not. Moreover, from “p” many different propositions could be generated, for example “not-not p” or “not-not-not p.” That is absurd, for “not-not p” and “not-not-not-not p” both are “p”. “Not” is not a function-name, but an operator.

Third, not only logical connectives such as “not” were misconstrued by Frege as function-names; expressions such as “object,” “relation,” or “concept” were misconstrued too. They were conceived by Frege and Russell as names of logical and ontological summa genera. On Russell’s view, they were the residue “from a process of generalization that has been carried out to its utmost limit” (Russell 1984, p. 98). Yet, Wittgenstein correctly argued, they are not names of entities of any kind (and are not part of logical propositions describing the most general forms of the universe, as Russell believed), but formal concepts and these are variables, not names. Because they are variables, they cannot be part of empirical propositions for these contain names of simple objects. And while we can observe, for example, a red object, there is no such thing as having “logical experience” (as Russell 1984, pp. 98–100, argued) by means of which we become acquainted with logical objects or relations. Moreover, we cannot give an experiential account of logical propositions such as “It is raining or it is not raining” or “If cows eat grass, then cows eat grass” in which logical operators (“or,” “if then”) occur, for experience cannot explain why they are necessarily true. If we learn something by experience, it is always conceivable that it may have been otherwise. Experience teaches us how things happen to be, not that things must be as they are.

Wittgenstein argued that Frege’s and Russell’s ideas lead to a misconception of empirical propositions. In Frege’s system, the function x2 would not be the function that it is, were its value for the argument 2 not 4. But this is problematic in the case of empirical propositions, for if we think Fa and it is true that Fa, then this is the same thought when we think Fa and it is false that Fa. A thought remains the same if it has a different truth-value. In Frege’s system, the sense of the empirical sentence Fa if it is true is different from the sense it would have if it were false. Hence Frege’s account of sense in his function-theoretic system, Wittgenstein concluded, is ill-suited for characterizing the essential nature of the empirical proposition.Footnote 6

Empirical, Logical, Mathematical, and Conceptual Propositions

Wittgenstein’s criticism of Frege’s and Russell’s ideas resulted in one of the most fundamental insights of the Tractatus. In opposition to Frege and Russell, Wittgenstein argued that empirical propositions (sentences with a sense) are not only bivalent (they are true or false) but essentially bipolar (they have a true and a false pole: they are capable of being true and capable of being false). Empirical, in contrast to logical (and mathematical and conceptual) propositions, can be true and can be false. We need experience to find out whether empirical propositions are true or false. Logical propositions such as “It is raining or it is not raining” are not bipolar and do not say what the case is. They combine empirical propositions in such a way that the resulting tautology is true no matter how things stand. They do not tell us anything about the world and are therefore senseless (they have zero sense). Yet the propositions of logic have a role in drawing valid inferences. That a certain proposition is a tautology, i.e., that it is well-formed but says nothing because it is so constructed that all content is cancelled out, can be said to be a rule. For example, to recognize that (x)(Fx)⊃Fa is a tautology (is true come what may) is to recognize a rule of valid inference. If someone recognizes this tautology, then they can draw valid inferences.

We can now also answer the question of how empirical propositions are distinguishable from mathematical and conceptual propositions (Wittgenstein 1978; Baker and Hacker 2009, Chap. 7). In contrast to empirical propositions, mathematical and conceptual propositions are not bipolar. Mathematical propositions or equations are rules for substituting and transforming empirical expressions concerning quantities and magnitudes. For example, if I have 25 bags with 25 apples in each, then the rule 25 × 25 = 625 allows me to conclude that I have 625 apples in total (without counting afresh). Hence true mathematical propositions correspond to valid transformations of empirical propositions about certain quantities or magnitudes of things. Equations also do service in generating other equations or theorems by means of proofs, resulting in a system of equations interwoven by networks of proofs. Negations of true mathematical propositions do not result in propositions that fail to correspond to the mathematical facts, for they are not like empirical propositions. Negations of true mathematical propositions are “stigmatized” as not having the standard role of mathematical propositions, that is, licensing the transformation of empirical propositions. Yet they may have a role within mathematics, for example in indirect proofs.

Conceptual (or what are called metaphysical) propositions are, just like mathematical propositions, part of a network, but they are not interwoven by proofs. In contrast to logical propositions, they are rules of grammar that are constitutive of the expressions used (for example, “Red is a color” partially determines how the word “red” is to be used). Importantly, they are expressions of rules in the guise of descriptions. “Red is a color” or “Red is darker than pink” look like empirical descriptions, but they are not. Their negation (e.g., “Red is not a color”) does not result in a false empirical proposition nor in a rule for the use of “red,” but in a nonsensical proposition. Hence conceptual propositions do not describe anything. They are norms or rules of representation. Some of them are rules allowing us to make inferences, for example if something is red, then “Red is a color” and “Red is darker than pink” license us (without looking) to say that it has a color and that it is darker than pink. That is what children learn when we teach them the network of color words.

The conclusion we can draw is that there are essential differences between empirical, mathematical, and conceptual propositions. Consequently, empirical, mathematical, and conceptual problems are resolved in different manners. The solution of mathematical problems requires new proofs resulting in a network of propositions connected through these proofs. Because mathematical propositions can be used to transform empirical propositions, we can apply them to phenomena resulting in empirical hypotheses that can be tested (a concrete example, concerning the application of Hamilton’s rule rb > c to empirical phenomena, is discussed in Smit 2014, pp. 92–93). Conceptual problems can be resolved by conceptual clarification. By resolving them, we remove incoherence in our network and can, from the new, coherent network, construct empirical hypotheses (examples, concerning the effects of imprinted genes on behavioral development, are discussed in Smit 2002, 2009, 2013).

The Pre-established Harmony Between Thought and Reality

Wittgenstein argued that an empirical proposition is bipolar and essentially connected to both truth-values. And since negation introduces nothing new into the proposition, it seems that what are true and what are false propositions is pre-established. This, what Wittgenstein (2005, section 43) called (with an allusion to Leibniz’s ideas) pre-established harmony between thought or proposition and reality, is a harmony that necessarily obtains no matter whether the proposition is true or false. Hence the harmony or agreement between thought and reality is not an agreement of truth. The question is then how we can understand this pre-established harmony between thought or language and reality. Wittgenstein’s answer to this question in the Tractatus is that propositions are logical representations or pictures. What makes it possible for a thought or proposition to constitute a representation or picture is that it is a picture that shares a common form with what it represents or depicts. We shall discuss some details of what others have called Wittgenstein’s picture theory, and also discuss why Wittgenstein later refuted this theory.

The connection of language to reality For understanding that thoughts or propositions are pictures with a logical form, it is useful to start with a discussion of the structure of propositions. Propositions are composed of expressions. The constituent expressions are analyzable (definable by analytic definition or paraphrase) or unanalyzable. The unanalyzable expressions are in Wittgenstein’s terminology of the Tractatus simple names and these link language to reality. For instance, we can link the names “tulip” and “purple” (in the proposition “This tulip is purple”) with objects and properties in reality if they are unanalyzable names.Footnote 7 These names have meaning: their meanings are the simple objects for which they stand.

Later Wittgenstein realized that he had confused simples (simple objects) with samples (Hacker 2001 Chap. 9; Baker and Hacker 2005). Take the example of the color red. For the explanation of the word “red” we can use a sample (e.g., a ripe tomato), since we can explain the meaning of “red” by pointing at the tomato. We do not connect then a name with a metaphysical entity (a simple), but use the tomato as a sample. When someone points at a tomato and utters the sentence “This is red,” they use this sentence together with the sample and the ostensive gesture as an ostensive definition.Footnote 8 It can be paraphrased as “This color is red.” The tomato is then used as a standard of comparison, which may be used to determine whether other objects in the surroundings are red as well, for they are correctly said to be red if they are the color pointed at. It follows that names do not connect language to reality, as Wittgenstein had thought in the Tractatus, for samples (standards for the correct explanation and use of words) belong to the method of representation (to what represents, not to what is represented). Explanations of names are therefore given within language. Moreover, Wittgenstein realized that this implies that correlating names with objects we see is not a mental act of meaning, but belongs to a linguistic praxis. The capacity to use names in accord with rules (such ostensive definitions) is a skill (or an acquired ability), not a mental act (see further below). Of course, that explanations of words are given within language does not mean that we cannot refer to objects in the outside word.

The harmony between language and reality We said that propositions have a structure. They do not contain logical connectives (such as and, or) and quantifiers (such as all, some). According to the Tractatus a proposition consists of a combination of simple names. We have seen that names have content (they stand for simples), but they also have an internal form. For instance, the internal form of a tangible object is that it is space-occupant: it cannot exist outside space (it exists in some location). The internal form of a property such as red is that it is a shade of color: the word “color” signifies the form of a visual object. That all color names (red, blue, etc.) are intersubstitutable salva significatione in well-formed sentences shows that they have the same form. And because they have the same internal form, the word “matte” or “glossy” can be associated with a color such as red, but “red” cannot be associated with “loud” or “soft,”—only sounds such as “G-flat” can be played loudly and softly (it does not belong to the nature of colors that they may be played loud or soft, just as red cannot be smelled). The internal forms determine how objects can occur in state of affairs. Hence the combinatorial possibilities that an object has is not a contingent matter (TLP 2.16–2.18). The totality of objects, with their internal forms, Wittgenstein called the substance of the world (TLP 2.021). It is what subsists independently of what is the case.

Besides internal properties objects have external properties. These are accidental, namely their concatenations with other objects resulting in states of affairs. A state of affairs is a possible combination of objects which may or may not obtain.Footnote 9 Empirical propositions have a sense: they describe a state of affairs in reality, namely that the meanings of their constituent names are combined in reality as the names are combined in the proposition.Footnote 10 A fact is a particular arrangement or configuration of objects, for example that this book is lying on the table (the book lying below the table would be another possible configuration). Statements of facts tell us how things are in the world. Thus, what an empirical proposition depicts is a possible configuration of simple objects (called a representing fact). An actual configuration of objects is a fact (the picture is then made true by the represented fact in reality).

I add two further clarifications. First, the elements of the picture need not resemble the elements in reality. This is obvious in the case of a linguistic representation. For example, the linguistic elements of “John loves Mary” do not resemble anything in the world to which they refer, whether they are spoken or written words. Second, just as the elements need not resemble the objects they stand for, so too the arrangement of the elements in a representation or model need not represent by resemblance. A musical score, for instance, may represent a piece of music, but the temporal succession of the notes is here represented by a spatial succession of the elements in the model or picture: the elements are arranged in a row from left to right.

Wittgenstein later repudiated his logico-pictorial explanation of the harmony between thought and reality. For an explanation it is useful to return to the problem of negation. We have seen that the early Wittgenstein argued that the thought that p is internally related to the fact that p that makes it true. And the thought that p is also internally related to the fact that not-p that makes it false. By internally related Wittgenstein meant that it is inconceivable that we think truly that p and it is not the case that p, and, likewise, that the thought that p would not be the thought it is, were it not the thought that is made false by its not being the case that p. Wittgenstein’s problem was how we can we understand these two truisms. On the one hand, it is clear that what we think (the picture or representing fact) cannot be identical or the same with what is the case (the represented fact in reality), for if it was, then we would think nothing when we think falsely that p (as Plato noticed). On the other hand, what we think when we think truly that p cannot also not be different from what we think when we think falsely that p, otherwise not-not-p would not be p. For this reason the early Wittgenstein argued that there must be something the same (the logico-pictorial form) and something different (the representing picture or fact is different from the represented fact) between what we think and what is the case.

The later Wittgenstein (1975, pp. 68–69; Hacker 2013a, Chap. 3; 2013b, Chap. 8) realized that he was misled here by grammar: the problem that was bothering him arose from asking the wrong questions. Take the example of identity. The grammatical point to notice is that the Wh-pronoun in “what we think” and “what is the case” is not a relative one. And because it is not a relative pronoun, the question, what is the same? may mislead us. For if we ask this question, we may, as the early Wittgenstein did, mistakenly presuppose that what is believed is some thing. But “what” is an interrogative rather than a relative pronoun. Hence, “what we believe” and “what is the case” are not two phrases in which the name of a thing occurs. With regard to things we can ask whether a thing is the same (or different) as another. But if “what we believe” and “what is the case” is the same, then both receive the same answer. We do not refer then to a tangible thing. This grammatical point clarifies why the early Wittgenstein was confused about how it is possible that what we believe can both be what is the case when the belief is true and not be what is the case when our belief is false.

The later Wittgenstein realized that he had misinterpreted a rule for the use of signs as a connection between language and reality. But there is no connection. It is true that one can read off the proposition that p the fact that makes it true or false, but that is not due to a logico-pictorial harmony between thought and reality, but is merely a move in grammar licensed by a substitution-rule: “the thought that p” = “the thought which the fact that p makes true” = “the thought made false by the fact that not p.” Similarly, an expectation or desire contains a picture of its own fulfilment, but that satisfaction is a logical relation, not a psychological one. An expectation or desire and its fulfilment make contact in grammar. It is not a relation between word and world; the relation is a grammatical nexus, a rule for the use of words: “the expectation that event e occurs” = “the expectation that is fulfilled by the occurrence of e” = “the expectation that is disappointed by the nonoccurrence of e.”

This grammatical resolution of the problem has consequences for how we should conceive of the content of our thoughts. When we believe that p (the content of our belief), that p is not some thing that is in our believing, let alone in our mind or brain. Hence when we say that it enjoys “intentional existence,” this is a misleading phrase, for it invites the unanswerable question of how we can explain it as a mental item. But it is not an item of any kind but is what specifies or individuates our belief irrespective of its truth. It does not describe a thing, but gives an answer to the question “What do you believe?” It is not an item in the mind, but a linguistic expression with a use. Similarly, facts, which are thought to be the worldly items in the world (matching our allegedly mental items), have no spatiotemporal location in the world either. They do not come into existence and then cease to be. The world does not consist of facts and facts are not true descriptions; rather a true description of some aspect of the world consists of a statement of facts.Footnote 11

Wittgenstein’s Early and Later Conception of the Intentionality of Thought

We have seen that, according to the early Wittgenstein, it is not sufficient that the assertion that p has something in common with what it represents. In order to represent the existence of a specific state of affairs, it must be meant and thought as a representation of it. When we think a possible configuration of objects, names are pinned to reality by acts of meaning so that the correlation between names and simple objects is effected. Mental acts of meaning, as it were, inject content into the constituent elements of a propositional sign. And by thinking the sense of a propositional sign the proposition is transformed into a representation (or model) of a certain situation: the proposition then depicts that situation (or represents the state of affairs) and expresses the thought that it obtains. The point to notice here is that meaning and thinking do not belong to the phenomena, for mere sounds cannot pin names to reality. Consequently, Wittgenstein’s early view results in a picture of the mental as something very special, since it can do things that mere sounds or inscriptions cannot do. Meaning and thinking are invoked to explain how life is infused into otherwise dead signs. As Wittgenstein (1958, p. 3) later put it:

So it seems that there are certain definite mental processes bound up with the working of language, processes through which alone language can function …the processes of understanding and meaning. The signs of our language seem dead without these mental processes; and it might seem that the only function of language is to induce these processes.

These ideas raise questions about how can we study these mysterious mental acts or processes which do not occur in the medium of sounds. Are meaning and thinking indeed mental acts, activities, or processes that infuse dead signs with life? When the later Wittgenstein realized that he was again misled by grammar, his answer was that they are not.

First, “to mean” is not an action-verb like “to say” or an activity-verb like “to converse” and “to discuss.” It does not make sense for one to intend or to decide to mean (we cannot be ordered to or agree to mean). Moreover, we cannot start to mean something, be engaged in meaning, and stop in the middle of it, as we can when we are engaged in an activity (meaning is also not a state or process that has genuine duration, with a beginning and ending). And, in contrast to an activity, we cannot mean something quickly or slowly as we can in the case of an activity. Some activities are more efficient or skillful than others, but we cannot mean something efficiently or skillfully. And meaning something is not an accompanying process or activity of saying something either. One can say, “While I was humming a tune, I was beating the time to it,” but not, “While I was saying that …, I meant ….” One can say, “When I said …, I meant …,” but this refers to the time of the utterance and not an activity that took place when one said something. Given these and other grammatical features of meaning the later Wittgenstein concluded it is not mental acts or activities of meaning that infuse life in dead signs.

Second, Wittgenstein also realized that there are no processes of thinking explaining how a proposition is transformed into a representation (or model) of a certain situation. Thoughts, Wittgenstein now saw, are not representations. A historical painting may be said to be a representation of the event it depicts, and a proposition (an assertoric sentence in use) may also be said to be a representation. If something is a representation, then it must have both representational features (pictorial features in the case of a painting, semantic features in the case of a proposition) and nonrepresentational features in virtue of which it can represent and be perceived to represent. The nonrepresentational features are the characteristics of the medium of representation. In the case of a painting, these are the oil, gouache, panel, and so on; in the case of a spoken or written proposition these are the timbre and loudness of the spoken voice or the ink with which we write a proposition. Thoughts, however, have no nonrepresentational features and are therefore not representations. In virtue of nonrepresentational features, we can perceive paintings and propositions, but thoughts, being imperceptible, do not have nonrepresentational features. They involve no medium of representation (they are all message and no medium). Of course, thought is expressed and communicated in a medium, but we do not think in a separate inner medium (in our mind or brain). There is no separate language of thought (ideas or concepts are not the constituents of thought) in which we think, although we can talk to ourselves in our imagination. Humans express and communicate what they think by answering the question: “What do you think?” But when answering this question, they do not translate mysterious mental symbols into a language (as, for example, Pinker 1995, p. 81, has argued), but invoke the rule-governed use of words and sentences of the language they have mastered for expressing their thoughts. The use of a sign and explanations of its rule-governed use is its life: it is our mastery of the techniques of employment. There is no hidden mechanism here in the mind or brain that enables signs to represent what they represent. Animals, by contrast, cannot do that, for they can only express their thoughts in nonverbal behavior. A dog may think that its master is going to take it for a walk when it hears its leash being taken from the peg. We say so because of its nonlinguistic behavior. But note that thinking (or knowing) that it will be taken for a walk is not the result of reasoning (inferring, deducing), but of perception and memory. Reasoning is restricted to a language-using creature who has mastered the use of logical connectives (“therefore,” “so,” “it follows”) and quantifiers, for it enables this creature to make chains of reasoning (to engage in thinking or reasoning).

How, then, should we characterize the thought “that things are so”? It is not a description of the thought we have when we think that things are so. Rather, it is a specification of the thought. It specifies what we think and does not describe what our thinking is like. And if we think that things are so and it is a fact that things are so, then the phrase “it is a fact that things are so” is not a description of a fact either. It is a statement of a fact. These claims of Wittgenstein are part of his later insight that grammar is autonomous: grammar cannot be justified by reference to facts, whether they are facts occurring in the physical or mental world.

Wittgenstein’s Alternative to the Ideas of Psychological Logicians

Wittgenstein’s criticism of Frege’s conception of logic has interesting consequences for how we should conceive of the laws of thought. We shall discuss in this and the next section Wittgenstein’s ideas and why he was critical of the ideas of the psychological logicians, and explain thereafter some consequences for evolutionary epistemology.

Scientists and philosophers have in the past discussed the question of how the laws of logic (like the law of the excluded middle) and rules of logical inferences are linked to actual thoughts, thinking, reasoning, and inferring. Are the laws of logic the laws of human thought? Are these laws of thought comparable to empirical laws and explicable in terms of our biology or psychology? How, then, can we understand that they are normative laws (telling how we must reason if we want to make a valid inference)? Before Wittgenstein started to think about these questions, answers were provided by the so-called psychological logicians (such as Erdmann). Psychological logic was a variant of empiricism that flourished in Germany in the 19th century. The aim of this tradition was to deepen logic by psychology. Psychological logicians thought that the subject matter of logic was inference, but argued that inferences were sequences of judgments. Judgments, i.e., propositions or thoughts, were thought to be built up out of concepts or ideas (these are the ultimate constituents of judgments). Ideas, in turn, are what is given in experience (I ignore here the alternative possibility, namely that ideas are innate, for this possibility is confronted by the same difficulties). They are the results of the impact of the material world, which consists in itself of matter in motion, on our receptive sense organs. For example, ideas of color, sound, or smell (sensible qualities) in the mind are generated by the impact of particles on our nerve endings.Footnote 12

This conception of the origin of ideas enabled the psychological logicians to explain how out of these simple ideas the mind weaves thought and language. According to them, thinking consists in combining and separating ideas by predicating or denying one or more ideas of other ideas. Language was explained by assuming that words are names of simple ideas in the mind. They assumed, just as the early Wittgenstein did, that simple ideas are not constructed by the mind but are indefinable and given in experience. Names of complex ideas are definable in terms of the names of simple ideas. Sentences consist of combinations of ideas: they are represented as conjoined by assertion, or separated by denial.

Wittgenstein argued that the doctrine of the psychological logicians is completely wrong. In this section we discuss why their ideas about concepts and sentences are misguided; in the next section we explain why they failed to explain the normativity of the laws of thought.

Mastery of the use of concept-words does not somehow flow from correlating concepts with subjective ideas (conceived as entities which are parts of judgments) and by abstracting from concrete experiences.Footnote 13 Children do not learn words and concepts (e.g., “red”) by abstracting from subjective perceptual experiences they have, but learn to use words and concepts when caregivers explain their use by, for instance, ostensive teaching. For example, caregivers teach children the meaning of “red” by referring to a ripe tomato in the outer world. The tomato pointed at, we have seen, is then used as a sample and can be used to determine whether other objects in the surroundings are red too. Notice that ostensive teaching is not merely ostensive training since it involves the normative explanation of linguistic expressions. The correct use of a linguistic expression is correct in so far as it accords with the rules for the correct employment of that expression. The capacity of children to use words in accord with norms or rules is visible in how they behave, i.e., in what they do or say. Whether a child possesses an ability is determined by testing whether he or she (1) uses a language (e.g., words) correctly, (2) explains its use correctly, and (3) responds appropriately to its use in context.

The technique of reasoning and thinking is learnt in a similar way. Children learn to reason and think when they learn to transform expressions according to a paradigm. For example, when we explain to children that lions are carnivores and ask them what Simba, a lion, will prefer (a piece of meat or an apple?), we teach them how to draw an inference (what correct inferences are). The point to notice is that rules of inference do not follow from the meaning of the logical connectives and quantifiers involved in inferences, but are constitutive of the meaning of logical connectives and quantifiers. Hence if we explain to a child by means of concrete examples that (fa) follows from (x)f(x), we give an explanation of what the universal quantifier means and teach him or her the technique of inferring (reasoning, thinking). Note that acquiring the ability to reason and think presupposes that children have already mastered some rules for the use of words and sentences. Hence the ability to reason evolves out of the ability to combine words and probably evolved during the course of evolution for similar reasons.

The Normativity of the Laws of Thought

We have seen that psychological logicians argued that judgments are combinations of ideas and inferences transitions from one judgment to another judgment. Consequently, the laws of logic or the laws of thought belong, according to them, to the domain of the mental and, hence, to empirical biology or psychology. The truths of logic have the status of empirical generalizations. Thus these propositions are not necessary truths, but rest on inductions based on introspection. This results in the radical view that they are a posteriori truths. Yet the laws of logic or thinking are not empirical generalizations but a priori truths. They prescribe how we ought to think. Psychologism in logic does not do justice to the distinction between actual and correct thinking. Logic articulates the laws of valid thinking or reasoning, whereas actual reasoning includes any reasoning by thinkers, valid or invalid. Investigating actual reasoning does not allow us to identify the subset of reasoning which is valid and does not help us to establish standards for classifying reasoning as valid or not. Take the law p.q⊃p.p, that is, if p, then no matter what else is the case (q), p holds. If p is “3 + 4 = 7,” and q “Harry is an ape,” then the law says that “3 + 4 = 7” holds. Now we can investigate people’s actual reasoning by asking what they think of the Harry-is-an-ape argument, but if someone argues that the reasoning is invalid, because Harry is not an ape, then we will conclude that this person has not yet understood valid reasoning (we presuppose here of course an understanding of the meaning of “⊃” (“if-then”) in the required sense). Hence the laws of logic tell us how we should think. Wittgenstein, we have seen, argued that they are rules of inference and are constitutive of what we call reasoning and thinking.

We have seen that the laws of logic and mathematics are not empirical or natural laws. Unsurprisingly, mutations do not alter the truth of laws of logic (or of mathematical propositions: 2 + 2 does not become 5 as the result of a change of our heredity). And there is not any reason to assume that experience may compel us to revise a law of thought. It follows that the connection between the laws of logic or rules for correct thinking and actual thinking is not that of a natural law and technical rules. Take the example of the connection between laws of metabolism and rules to prevent pathology. For example, eating citrus fruits to prevent scurvy is an efficient means to achieve a goal (prevent disease). This rule can be understood independently of the laws of metabolism (assuming that there is a law describing a biochemical mechanism in which vitamin C has a role). Frege once argued for this reason that a similar connection exists between laws of logic and rules for correct thinking. But the laws of metabolism are empirical laws, whereas the laws of logic are normative laws and these are not independent of the concept of valid inference. The laws of logic are internally related to correct thinking.Footnote 14 By comparing the laws of logic to the laws of nature (metabolism), Frege hid from view the fact that the propositions of logic have a completely different role in language. They are, as Wittgenstein showed (TLP 5.2–5.24), not the most general laws of thought, but tautologies, i.e., propositions with zero sense which say nothing at all. And because they are tautologies, they cannot function as grounds justifying technical rules which regulate human thinking (the propositions of logic do not have content, as was assumed by psychological logicians). Rather, the conceptual connections between the tautologies and rules of inference is that to acknowledge that a pattern of propositions exemplifies a rule of inference, for example, that “p, if p then q, so q,” is to acknowledge that the corresponding conditional expresses a tautology, i.e., that the proposition “p.p⊃q.q” is a tautology. Proofs of a proposition of logic do not establish the truth of one proposition from the truth of a previously established proposition. They only show that a certain formula is a tautology: a “proof” shows that a proposition of logic is a tautology because it can be derived by transformations from other formulas that are tautologies.

Wittgenstein’s ideas raise a question. If the laws of logic and inference rules are constitutive of thinking, what would it be like if someone violates our laws of thinking? Suppose that beings bring off judgments contradicting the law of identity. They say, for example, that “a is not identical to a.” When we are confronted by a creature uttering this sentence, we would probably at first assume that they do not know what we mean by identity. But suppose that it is not a matter of misunderstanding, then we may wonder what they think when they express thoughts involving the concept of identity. Can they, say, express the thought that this book is identical to itself? Can “this book is not identical to itself” express a different thought than the thought we express when we say that “this book is identical to itself”? The problem we encounter here is akin to describing what it would be like to move rooks diagonally in chess, or to castle in draughts. The rules of chess show us how to play chess, and when we violate these rules, we no longer play chess. Similarly, when we encounter someone who apparently reasons according to different laws of logic, then it is ill advised to try to explain logical impossibility in terms of a further impossibility. It is better to argue that we encounter not a different kind of reasoning, but something that is not reasoning at all (a “speaking with tongues”), for the laws of logic are constitutive of what we call thinking. The grammatical impossibility here is that it does not make sense to assume that beings who violate the laws of logic can be said to be engaged in what we call “thinking.” Consequently, it does not make sense to say, as Frege did, that our logical system is superior to their system (although we can, of course, argue that certain rules are more appropriate for certain purposes than others). Suppose there are beings whose laws of thinking are contradicting our laws; what would we say? Frege (1964, p. 14) said that we have here a hitherto unknown kind of madness. It makes sense, according to Frege, to say that there are beings who think according to different laws of thought, but he thought that these beings are wrong and we right for only our laws are in accordance with the laws of truth. However, as Wittgenstein (1978, I, par. 152) correctly remarked, the problem with Frege’s response is that he never said what this insanity would be like. Rules of inference are constitutive of thinking and reasoning. Consequently, beings that follow different rules than the ones we follow are not engaged in reasoning.

Kant, Wittgenstein, and Evolutionary Epistemology

Descartes, and in his footsteps many others, argued that we have direct, privileged access to our subjective experiences. He assumed that the primacy lies in statements of judgments concerning these bare experiences while statements of judgments concerning objects were supposed to be secondary to judgments of experiences or thoughts. His ideas created a gap between descriptions of subjective experiences in the “internal world” and descriptions of objects in the “external world.”

Kant criticized the idea that our knowledge is inferred from subjective experiences. One argument was that metaphysical truths (for example that every event must have a cause, or that substance must persist throughout change) cannot be derived from experience, for experience can yield only contingent truths, while the truths of metaphysics are according to Kant both universal and necessary. The truths of metaphysics, Kant argued, are known a priori. But how can we know these truths? Kant’s answer was that metaphysical knowledge is neither analytic nor empirical. Metaphysical truths are synthetic a priori truths. Our knowledge of these truths does not have to conform to objects, as idealists assumed. Objects, insofar as we have synthetic a priori knowledge of them, have, according to Kant, to conform to the a priori condition of the very possibility of experience (resulting in empirical cognition). Thus, the mind imposes structural principles on nature as a condition of possible experience. As Kant ([1787]1929, Critique of Pure Reason, B xxiii) put it: “Nothing in a priori knowledge can be ascribed to objects save what the thinking subject derives from itself,” for “we can know a priori of things only what we ourselves put into them” ([1787]1929, B xviii).

Kant’s ideas raised among biologists and philosophers interested in evolution the question whether we can understand the a priori conditions in evolutionary terms. The basic idea of evolutionary epistemologists (for example, the Austrians Lorenz (1941) and Popper (1972)) is that the conditions (the categories of the understanding such as space) are a priori to the individual (or individual development), but a posteriori to the species to which the individual belongs. They assumed that during the course of evolution, species, by adapting to environmental conditions, acquire information about the environment (stored in DNA). This phylogenetically acquired information, when passed on to future generations, is then a priori to the individual and structures its experiences. Hence, Popper (1972, p. 69, footnote 31) concluded that, “Kant was right that it is our intellect which imposes its laws—its ideas, rules—upon the inarticulate mass of our ‘sensation’ and thereby brings order into them.” We shall argue that Kant was wrong because he mistakenly still adhered to some presuppositions of the subjective conception of representative idealism. We will explain this by discussing some essential differences between Wittgenstein’s and Kant’s ideas.

Like Kant, Wittgenstein repudiated foundationalist epistemology of Descartes and Locke. Our knowledge of how things are in the world is not inferred from how things sensibly seem to us to be, as rationalists and empiricists believed. And, just as Kant did, Wittgenstein rejected Cartesian and empiricist conceptions of the mind: the mind is neither an immaterial substance as Descartes thought, nor, as for example Hume thought, a bundle of perceptions. Yet there is an important difference between Kant and Wittgenstein. In opposition to representative idealism (of Descartes and empiricists like Locke), Kant argued that metaphysics should study the—what he called—transcendental roots of the possibility of conceptualized experience. Wittgenstein also repudiated foundationalist epistemology and rejected Cartesian and empiricist conceptions of the mind too. Given these similarities with Kant’s ideas, it has been believed in the past that Wittgenstein advanced a logical form of transcendental idealism. Whereas Kant restricted the categories mainly to the concepts of modern physics (time, space, cause, and so on), it was thought that Wittgenstein expanded the forms of thought or reasoning with the concepts of modern logic (pioneered by Frege and Russell). Yet we know now that, despite some similarities, Wittgenstein did not advance a form of transcendental idealism. For explaining the difference between Kant and Wittgenstein it is essential to understand what is meant with transcendental idealism and how Wittgenstein construed his alternative ideas.

Transcendental idealism is based on the distinction between things as they are in themselves and as they appear to us. The world exists (in itself) independently of us, but the way it appears to us is determined by our cognitive faculties. In Kant’s philosophy, all objects are experienced in space and time and according to causal laws, but this is not due to their independent nature (as they are in themselves), but is imposed on the world by the perceiving subject. Hence Kant’s idea that the conditions of the possibility of empirical knowledge have transcendental idealist roots: categories like space, time, or quantity are not derived from experience (as representative idealism supposed), but are a precondition of it. Although it has been suggested in the past that logic in the Tractatus has the transcendental role that space, time, and so on have in Kant’s philosophy, that suggestion is misleading. For logic and language do not shape the world (or produce the form of the world), but, according to the early Wittgenstein, mirror or reflect the world. The later Wittgenstein realized that his early logico-pictorial explanation of the harmony between language and reality was misguided. It is true that one can read off a proposition that p the fact that makes it true, but that is not due to a metaphysical harmony between language and reality, but is merely a move in grammar licensed by a substitution-rule: “the proposition that p” = “the proposition which the fact that p makes true.” The later Wittgenstein concluded that the conditions of the possibility of empirical knowledge are norms of representation and these are not language-independent (they are explained by intra-grammatical elucidations). Thus, Wittgenstein’s conception of logic and metaphysics is not subjective in the transcendental sense (as opposed to what Kant and later Popper and Lorenz argued): it is not imposed upon the world by us. The intentionality of thought, the later Wittgenstein saw, is parasitic on the intentionality of language and linguistic representation. Thoughts, we have seen, are not representations. Consequently, current accounts of biologists and philosophers of biology in which it is argued that we need to invoke mental representations for understanding how cognitions and language can represent reality are on the wrong track. Unsurprisingly, Wittgenstein’s ideas result in a conception of the mind that radically differs from the one found in current biology and psychology. The Wittgensteinian (also called neo-Aristotelian) conception of the mind and how we can use it to construct explanations of the development and evolution of the mind has been discussed elsewhere at length (see, for example, Smit 2014, 2018, 2020).

Expressions of Experiences are Groundless

The second fundamental problem concerns the grounds of experiences. This problem is again a Cartesian legacy. According to Descartes, to have a mind is to have experience and to be aware of oneself as having experience. Yet how do we know how things are with us? Is it by inner sense or introspection? Kant had already criticized Cartesian and empiricist variations of representative idealism. He argued that it was an error to suppose that we know how things are with us by means of inner sense. That we can say how things are with us is not grounded in introspection. Kant argued that all that is required is that it must be possible for me to be conscious of my experiences (perceptions, desires, and so on). Only then can I represent the sensible experience as mine. The point to notice is that Kant left here a small part of the Cartesian and empiricist conception intact, namely that it is possible for the subject of experience to be conscious of experiences owned by themselves and, hence, to know that things are thus-or-so with the subject. Kant captured this by saying that it must be possible for the “I can think” (or “I can see”) to accompany all experiences (or representations). Notice that this is not an epistemic requirement (as Cartesians and empiricists thought), but a formal one.

Yet Kant still was confused for he discussed issues in terms of Descartes’s misguided conception of consciousness. Although Kant realized that conceptualized experiences are not self-ascribed to a separate, mental substance or self ([1787]1929, Critique of Pure Reason, A 355), he retained the Cartesian idea of the self-ascribability of conceptualized experience (or representations): my being conscious of experiences is self-ascribable to the transcendental subject without explaining what this transcendental subject is (see Hacker 2013b, Chaps. 2 and 3). Wittgenstein went further: we can do without Kant’s formal requirement. It is sufficient to argue that the “I can say” (not the “I can think”) accompanies my representations, for as language users humans can express what they think in a language (and that is a skill they acquire). The difference with Kant is that Wittgenstein ([1953]2009, par. 243–315; Hacker 2019) argued that there are no grounds for my being able to say what I experience, feel, think, or desire. We can say that we are in pain, that we see or hear something, that we think that things are thus-and-so or that we intend to do such-and-such without any grounds. For a brief explanation, we return to the bipolarity of empirical propositions.

“I am in pain,” just as “He is in pain,” is an empirical proposition: it can be true and can be false. The proposition may be used as an avowal or report. Like other contingent empirical propositions, its truth excludes a possibility, namely my not being in pain. Both the proposition and its negation make sense and describe real possibilities. Now suppose that we ask the traditional epistemological question of how I know that I am in pain. Can we give an empirical answer to this question? For answering this question, we can ask whether “I know that I am in pain” is also an empirical proposition that can be true and can be false. Does “I know that I am in pain” exclude a possibility, namely its negation? The answer is no, because “I don’t know that I am in pain” is nonsensical, for it is equivalent to “I am in pain and I don’t know it” (which is akin to “It is raining but I don’t believe it is raining”) and that is a patently nonsensical sentence.

If we assume that to say that “I know that things are so” implies an empirical issue, then we suppose that there is a way of finding out whether things are so. But there is no such thing as my finding out whether I am in pain, as opposed to finding that I am in pain. And there is also no such thing as learning or coming to know that I am in pain. In general, if we know that things are so, then we must be able to answer the question “How do you know” either by citing a source of knowledge (hearing, seeing, and so on) or by citing evidence or hearsay. But although we feel pain, “feeling” here does not signify a source of knowledge (while tactile perception, e.g., feeling with one’s fingers whether something is rough or smooth, is a source of knowledge). And because there is no such thing as coming to know whether we are in pain, it also makes no sense to talk of guessing, surmising, and conjecturing whether things are so. Consequently, justifying, conforming or disconfirming, and ascertaining are excluded too.

We can conclude that there is no epistemic work for “I know” to do, nor do we need to invoke Kant’s transcendental idealistic account of the “I can think.” The long tradition in epistemology correctly argued that “I do not know that or whether I am in pain” is excluded, but wrongly concluded that it is excluded for epistemic or transcendental reasons (for further discussion, see White 1982; Hacker 2013a, Chaps. 4–6). It is excluded for grammatical reasons (it makes no sense to say that “I do not know that or whether I am in pain”). And since we are not ignorant for epistemic reasons, there is no such thing as being in pain and knowing it, for we have seen that there is nothing for “I know that or whether I am in pain” to exclude.

Wittgenstein’s criticism of representative and transcendental idealism raises a question: if first-person present tense expressions are not grounded in subjective experiences, what, then, is the correct alternative to the Cartesian idea that our knowledge of objects and their properties in the external world is secondary to our knowledge of subjective experiences? The answer is that, in opposition to what Descartes thought, learning the use of the names of objects and their properties precedes the acquisition of perceptual verbs (for perceptual verbs are operators on nominals specifying an object of perception). Only if children understand the use of words such as “rose” and “red,” can they learn to say sentences such as “I see a red rose.” Perceptual verbs are being mastered when children learn to respond to questions such as “Can you see …?” or “Did you hear …?” and to instructions such as “Look! What do you see?” And when they learn to use these verbs by answering such questions, they also learn to answer the question “How do you know?” For example, if a caregiver asks a girl, “How do you know that Mummy is in the garden?” she can answer that she saw her there. By engaging in these language games, they begin to understand that the exercise of the senses is properly cited as a source of knowledge. They then understand that when someone sees a red rose and says so, the speaker is introducing a report on public visibilia (“I see …”) or justifying such a report by citing what he or she saw. The utterance is not grounded in subjective experiences, as Descartes thought, but specifies the sense-faculty by the use of which they came to know things to be so (what they say to be so). Hence by affixing the operator “I see” to “that things are so,” the exercise of the senses is cited as a source of knowledge.

When children are mastering the use of perceptual verbs, they not only learn that the senses are a source of knowledge, but also that observation claims are defeasible, for observation conditions are sometimes suboptimal, sense organs are sometimes defective (temporarily or permanently), and the objects may look other than they are. When they make errors, their parents will correct these and will explain how they can correct errors: by looking again or by improving the observation conditions (moving closer, turning on the light). They then learn to use qualifying expressions such as “This looks like …” or “I think it is …,” and begin to understand that these operators are used when the employment of the senses and the cognitive faculty was not optimal, enabling them to understand error and illusion.

We can conclude the self-predication of perceptual verbs and their cognates is not grounded in subjective experience, but rests on perceiving what is in one’s field of perception (what we see with our eyes in the external world). The groundless application of verbs of perception is learnt as operators on descriptions of perceptibilia and as indicative of validating sources of knowledge.

There is one question remaining. How should we conceive of the third-person present tense propositions such as “He is in pain” or “She sees a red rose”? Wittgenstein’s answer is that it is an intrinsic feature of mental concepts that they are Janus-faced, that one can be said to possess them only if one has mastered the grounds for applying them to others and can apply them groundlessly to oneself. We teach children the use of the psychological vocabulary in this domain by getting them to apply to others on the grounds of behavioral criteria and by getting them to use these expressions in first-person avowals and manifestations of experience. These are two sides of one and the same coin. Hence mastery of psychological predicates begins when children learn to apply them groundlessly to themselves and to others on the grounds of behavioral criteria. We have argued that we can explain the evolution of this capacity in a similar way (Smit 2016, 2018, 2020). We have also explained that, when humans could recollect their doings and reflect on them, it can be said that they had personal experiences (Smit 2022). They could explain these experiences to others and clarify why they did something (their motives) by reference to reason-giving explanations. The evolution of these complex linguistic skills led to what Aristotle called our second nature. It freed humans from the constraints imposed by genetic evolution.

Conclusion

Lorenz (1941, 1965) suggested that, after Darwin, it must be possible to explain the a priori condition of the very possibility of experience in terms of the brain. For we know now that the brain is an organ that evolved during the course of evolution (however, notice that Lorenz never gave a concrete evolutionary explanation). Lorenz noted that the only alternative is to assume a pre-established harmony between thought and reality, and added that this is not an explanation and does not provide us with any further insights.

Lorenz’s suggestion seems reasonable as long as we assume that we are dealing here with empirically solvable problems. That is, however, a dubious assumption. Kant and later Wittgenstein would have been surprised to read that they were investigating empirical problems (instead of conceptual or metaphysical problems whose resolutions require a priori arguments). We have discussed that, when Wittgenstein was investigating the pre-established harmony between thought and reality, he investigated the conditions of the possibility of representation in terms of an agreement of form (a harmony) between a thought and what it depicts irrespective of whether it is true or false. It did not concern the question of how we know whether a thought is true or false, but the undeniable internal relation between a thought and what makes it true or false, between an expectation and what fulfils or disappoints it, and so on. By discussing the a priori resolutions of the two fundamental problems, the later Wittgenstein has provided us with a sound conceptual alternative to representative and transcendental idealism. However, as we said in the introduction, that alternative is revolutionary and therefore difficult to grasp. A main difficulty is that Wittgenstein’s repudiation of representative and transcendental idealism concerns not a few ideas but a large battery. Wittgenstein argued that the conceptual framework of representative and transcendental idealism, in which definitions, propositions, inferences, and so on seem to mutually support each other, should be replaced by another framework, for the framework consists of a whole battery of mutually supporting incoherences. And that implies that hypotheses generated within the confines of the incoherent framework are not empirically testable hypotheses, as they are not hypotheses at all. Since it involves a conceptual revolution, it requires no stretch of the imagination to see why Wittgenstein’s insights are so difficult to understand and to accept. If we take them seriously, we should stop asking unanswerable questions and should pose different ones.Footnote 15 Thus, in contrast to what Lorenz thought, what can be explained is not how the a priori categories of understanding are encoded in the brain or DNA, for we have seen that they are explicable in terms of norms of representation. And norms of representation are rules of grammar belonging to the representation method. And, in contrast to what current biologists and philosophers of biology argue, what we have to investigate is not how expressions like “I think” or “I am conscious” are rooted in introspection, apperception, or feelings, but how linguistic forms of life evolved as extensions and replacements of animal forms of life. In neo-Aristotelian terms: how did the human rational psuchē evolve out of the sensitive psuchē of the other animals? That requires us to study how the ability to use linguistic expressions (including “I think that …” and “I am conscious of …”) emerged during the course of evolution and human history (and nowadays emerges when children learn to participate in the normative practice of using words and later sentences), for doing things with words led (and leads) to an expansion of our mental powers. That is a research project that I, as biologist, have found (and still find) an exciting challenge, although it required a profound change of my way of thinking.