---

TOP UPDATES FOUR PILLARS CINEMA/TV GAMES MANGA/ANIME MUSIC WRITINGS FAQ LINKS

LUDWIG WITTGENSTEIN EXPLAINED: PHILOSOPHICAL INVESTIGATIONS (1953)

Table of Contents

1. Introduction
2. The "Later" Wittgenstein
1. Ludwig Wittgenstein's Life, Post-1921
2. The Nature of the Transition From "Early" to "Later"
3. Structure of the Philosophical Investigations
4. Summary of the Philosophical Investigations
1. §1-37: Language-Games and Meaning as Use
2. §38-64: Naming Objects
3. §65-88: Family Resemblances
4. §89-137: Philosophy as Therapy
5. §138-242: Understanding and Rule-Following
6. §243-275: The Private Language Argument
7. §276-315: The Beetle in the Box
8. §316-362: Thinking and Talking to Oneself
9. §363-397: Mental Images and Imagination
10. §398-427: The Self and Consciousness
11. §428-465: Intentionality
12. §466-490: The Senselessness of Skepticism
13. §491-570: Nonsensical Sentences
14. §571-610: Psychological Experiences of Qualia
15. §611-628: Free Will
16. §629-660: Intention
17. §661-693: The Meaning of "Meaning"
18. Part II: Seeing an Aspect

1. Introduction

I have published an explanation/commentary on the first of Wittgenstein's two major works, the Tractatus Logico-Philosophicus (1921). This seeks to be a similar treatment of his later work, the Philosophical Investigations. I don't have a visual novel to thank for introducing me to this one like I do for the Tractatus, but it is a remarkable work of philosophy all the way. The Tractatus lecture is not strictly necessary to read before this one, but I recommend doing so for a background of the differences, real or imaginary, between the "early" and "later" Wittgenstein.

2. The "Later" Wittgenstein

2.a. Ludwig Wittgenstein's Life, Post-1921

Ludwig Wittgenstein published the Tractattus Logico-Pilosophicus in 1921. He was convinced that he had solved all philosophical problems in it and was thus done with philosophical work. After this, he retreated from public life. He gave away the majority of his immense familial wealth to various artistic organizations that he supported and spent almost a decade working at elementary schools in rural Austria. His return to philosophy was in 1929. He returned to the University of Cambridge and became a lecturer, and a very idiosyncratic one. His lectures apparently often felt more like watching him think through philosophical problems alone with the students being more like an audience that just happened to be observing.

From 1936-1938 Wittgenstein again retreated from public life and spent a lot of this time living in isolation in a small cabin in the remote countryside of Norway. He wrote most of the material that would evolve into the Philosophical Investigations while living in these conditions in sparan austerity.

In 1938, Nazi Germany annexed Austria. The Wittgenstein family had Jewish grandparents but had converted to Catholicism and had severed just about all Jewish cultural influences in their children's lives. His family were nevertheless designated with "Mischling" status, meaning they were considered mongrels who were not fully Jewish or Aryan. Wittgenstein became an English citizen and most of his family followed suit by fleeing for England or the US. He never returned to Austria.

Wittgenstein spent the rest of his days lecturing and working on manuscripts that largely remained private. He often seemed very distant even to those who knew him closest, but he regardless had quite a cultish following of passionately devoted students and friends. Wittgenstein was diagnosed with prostate cancer and died in 1951 at the age of 62.

2.b. The Nature of the Transition From "Early" to "Late"

The Philosophicus Investigations was published in 1953. Like all of Wittgenstein's later works, it cannot be considered a complete text that appears in the form Wittgenstein ultimately would have wanted. His notorious perfectionism ensured that only the Tractatus would ever reach this status. But we do have a substantial enough selection of writings to make some clear statements about his later thoughts. Now we can ask: If Wittgenstein considered himself to have solved all philosophical problems in the Tractatus, what could have changed his mind? How did the later Wittgenstein evolve and revise the work of the early Wittgenstein?

I can't sum up the entire Tractatus in a paragraph. I wrote a lecture about it. This lecture will probably be more comprehensible if you read that first, but it's not strictly necessary. The essential charge of the Tractatus seemed to be that all propositions, including those of philosophy, gain meaning insofar as they are pictures of states of affairs that may or may not be true in the world. Insofar as they have meaning and allow communication and assessment, they share a logical form in common with the states of affairs they stand for. Much of what goes on in philosophy is truly meaningless, because certain signs in the propositions philosophers use have no meaning or reference and therefore convey nothing.

The Tractatus is a highly "destructive" work. It is meant to end further philosophical work rather than contribute towards it, at least under our traditional conception of philosophy. In a sense, Wittgenstein's later work continues down this road. However, perhaps the best way to understand the fundamental change is to recount an experience that Wittgenstein cited as being very educational. It occurred when he was talking to the economist Piero Sraffa, who had also moved to Cambridge after the fascist takeover of Italy. Wittgenstein explained his idea of propositions conveying meaning by sharing a form with what the pictured. Sraffa replied by brushing his fingertips outward from under his chin (a gesture which is essentially the Neapolitan form of a middle finger) and playfully asked: "What is the logical form of THAT?"

Wittgenstein's later work becomes further detached from the quest for logically perfect language that motivated Russell and Frege and goes in a radically different direction. His work continued to try to point out what was nonsense and get rid of pseudo-problems of philosophy. But under the influence of encounters like those with Sraffa, he no longer believed that there was ONE logically perfect way to show the essence and meaning of any propositions. He now believed that all meaning and comprehension occurred within what he called our "stream of life."

The Tractatus Logico-Philosophicus makes the following assertion near its end:

This is essentially what Wittgenstein attempts to do in the Philosophical Investigations. But now he has found a way to do so in a way which is not "unsatisfying" like he claimed it would be. He does so in a way that is actually very therapeutic, enlightening, and constructive towards clear, fruitful thought. He wants to do so in a way that frees us of all the woes of philosophical angst and show us that, in a sense, everything before us is already known and already perfectly in order.

3. Structure of the Philosophical Investigations

Before I get into directly taking the reader through the text of the Philosophical Investigations, I need to explain the structure and method of the book a bit. The work is unfinished and it shows its origin as a sort of manuscript of notes and exercises. It consists of 693 short numbered sections which range in length from a few words at the shortest to about a page or two at the longest. There are no chapter breaks or broader groupings of these sections. They clearly move through different arguments and themes, but where exactly to put bounds aruoud the different drifts of thought is conventional and by no means clearly decided.

The Tractatus Logico-Philosophicus appeared as a number of austere philosophical axioms. It appeared almost more like a mathematical proof than a philosophical essay. The structure of Philosophical Investigations is completely different but no less unfamiliar. Most of the sections appear as short aphorisms and they are often quite cryptic. In most cases, they put forward questions witohut explicitly answering them, trying to get us to fill in the gaps.

The work appears as a dialogue of sorts at times, because there's often a so-called "interlocutor" who asks critical questions which appear in quotation marks. The work could be compared to a Socratic dialogue, but perhaps a better analogue would be the record of a Zen master's kouans, as the answers he gives rarely lead to an answer as thoroughly and systematically as Plato. Sometimes the answers only make us more confused!

As such, this is a very difficult work to understand, but a lot of it is simply because speaking in traditional terms of argumnet and thesis starts to seem inadequate. But there is much to learn regardless and it is quite a consequential work for all kinds of disciplines.

I am deeply indebted to the immense, thorough, comprehensive exegesis provided by Gordon P. Baker and P.M.S. Hacker. I've used their four-volume series to supplement and augment a more general overview by Marie McGinn in the Routledge Philosophy Guidebook series which is very approachable but omits large sections of the text. I will do my best to add citations where I've paraphrased them most heavily. All quotations are from the 2009 Wiley-Blackwell edition with a translation by G.E.M. Anscombe, P.M.S. Hacker, and Joachim Schulte unless otherwise noted.

4. Summary of the Philosophical Investigations

4.a. §1-37: Language-Games and Meaning as Use

The Philosophical Investigations opens with a passage from the first book of Saint Augustine's Confessions where Augustine begins to tell his life story by sketching out how he learned language and came to be able to express himself:

Wittgenstein's text opens with an extended critical analysis of the picture of language that Augustine sketches out here. Of course, Augustine never wrote this passage in order to put forward a "theory" of language. He was just describing his upbringing. But that is precisely why it is so valuable. It is an example of how we unthinkingly tend to conceptualize language. Many of us would probably put forward a similar picture.

Of many important implications within it, the first to address is the way Augustine seems to take words as "names" for objects and sentences to be combinations of these names. Augustine is probably thinking of one particular sort of word here, namely nouns like table, chair, bread, and peoples' names. These are the types of words that we learn in the manner that Augustine describes. That is, in an "ostensive" fashion. Ostensive is a fancy word that gets used a lot in this text and its meaning is a "direct" demonstration. If we define something in an ostensive fashion, we do so by "pointing" to it other otherwise directly indicating it. An ostensive definition essentially takes the form "THIS is a table" instead of trying to define "table" by saying "a wooden slab with four legs to set things on" or whatever other dictionary definition we would give it.

And of course it makes sense that Augustine thinks that the essential grounds of language amount to this ostensive kind of "pointing." Clearly we know that when we teach babies to use language we do a lot more ostensive pointing than reading them definitions out of the dictionary. But Wittgenstein says that this is at best an incomplete view of language. We can't look at the way we master only one type of word and base our entire conception of language on it. He says that this is like saying that playing a "game" consists in moving objects around on a surface according to certain rules. That sounds like a description of a BOARD game. But there are many other types of games that this definition leaves out.

Wittgenstein now images what a language that is in fact accurately described by Augustine would look like. This would be a language where all words are names for objects. This language exists between a builder and his assistant and, for the sake of simplicity, will contain only four words referring to the different kinds of stones they need to build something: "block," "pillar," "slab," and "beam." When the builder needs a certain kind of stone, he calls out, for example, "Slab!," and the assistant hands him the appropriate one as he's been taught to. Now imagine that this language was the entire language of a tribe. Children in that tribe will be brought up to hand slabs over when hearing "Slab!" and so on. How do these children come to learn language?

At this point we might want to say that the children have a full mastery of the language when they come to make a mental association between the shapes of stones and the respective words. However, Wittgenstein thinks that even in such a primitive language, this isn't really the purpose of these words for the tribe. We could imagine a language whose purpose was just to make a certain object come before our minds when we hear a word. In his words, it would be a language where speaking a word was like striking a note on the keyboard of the imagination. But this is certianly not the purpose of these words for the tribe in question. They need these words in order to accomplish the act of building. It may even be that having the proper stone appear in the mind as an image when the word is spoken could help the goal of building. But it is not the goal itself.

We have the same tendency as Augustine when we try to explain what a word means by focusing on nothing but the mind of the person who hears it. This might give us the illusion that we're getting closer to the "real" core of what the meaning of a word consists of. But we are actually moving away from it rather than towards it, because we are removing all the background contexts of use that really give words their meaning. The fact is that language is never a free-floating structure that we can isolate to understand in full. In fact, Wittgenstein says that everything we need to know to explain language already lies open to view. Language only ever gains meaning in the full context of activities into which it is woven and in which it emerges. Wittgenstein calls these "language-games." The full context of interactions in the tribe of builders is one such language-game. [2]

Let's imagine that the language of the builders becomes more complex. Let's say that they add the following:
1. number words to count stones (represented by the letters a, b, c, d)
2. words which are the equivalent of "this" and "there" that are used along with pointing-gestures
3. a number of physical color samples
Now a builder could give an order like "d-slab-there" while holding up a color sample to show what color slabs he wants and pointing in a certain direction. These new classes of words already function quite differently to "slab" and "block." How do we teach these words in an ostensive fashion? We could possibly teach numbers for example by counting a, b, c slabs with the appropriate number of slabs. But "this" and "there?" If we point at a table and say "this," how do we know that "this" is not the word for "table?" When we point and say "there," how do we know that "there" is not the word for a specific place like a hut which lies in that direction?

We may want to ask what these words "signify," but this is where we get misled by strange philosophical instincts. How do we know what words signify other than how they are used? And we already have all that. We have the desire to make some canonical statement of "how" all words can signify. We want something that includes the way different words like "block," "d," and "this" all signify. But this is the mistaken instinct which Wittgenstein calls the philosopher's "craving for generality."

Why do we assume that being more general and creating an overarching system is always the direction towards greater understanding? In this case, it obscures the real use and meaning of language rather than clarifying it. Wittgenstein uses the analogy of tools in a toolbox. There are many different types of tools in a toolbox like a screwdriver, a ruler, glue, nails, and so on. There is nothing we can say about the general way that all these tools are used other than the fact that they are indeed used. And it is just the same for words. We are only confused because words have a more uniform appearance in print or speech.

We think of examples like "block" and "slab" and imagine that all ways of signifying are like attaching a name tag to an object. But clearly words like "this" or "d" do not work like this. In fact, on what account can we even say that the color samples are not part of the language of the builders? They are not spoken utterances, but we certainly cannot find any difference between them and the spoken words insofar as they serve a use in the language-game.

Here we have the essential difference between Augustine and Wittgenstein. Wittgenstein is not just trying to enlarge the taxonomy of Augustine's concception of language. That would involve moving AWAY from language and isolating it from how it is used in the "stream of life." This goes completely against the spirit of Wittgenstein's investigation.

We might be concerned that the language of the builders can't be compared to ours because it is not "complete." But a language can only be considered "complete" if we conceive of language as having some central essence that involves modeling the structure of reality. This was the picture that Wittgenstein seemed to put forward in the Tractatus Logico-Philosophicus, but now the conception has changed. Language does not have any ONE essence. Language is a number of different techniques that we use in various language-games.

Wittgenstein claims that "to imagine a language is to imagine a form of life" (§19). This term, "form of life," is the most important one in this section and perhaps in the text as a whole. Language is always embedded in a "form of life," that is, a background of non-linguistic behavior. Forms of life shift, go extinct, get born, intersect with each other, and so on. The idea of a "complete" language ceases to make sense when we understand this. It makes us much sense as a "complete" tool kit. In the Tractatus Logico-Philosophicus, one proposition claimed that "The limits of my language mean the limits of my world." In a sense, this is still THE motto of Wittgenstein's philosophy. But now this limit is not a sort of definite and ordained boundary set up by logic. It is a fluid and evolving structure that exists within the stream of human existence.

When one of the builders says "Slab!" in his very primitive language, is he saying a "word" or a "sentence?" If it is used as a word, then it certainly can't have the same meaning as the word "slab" in our language, as for him "Slab!" is a complete command. We might be tempted to say that his "Slab!" is really an abbreviated version of the sentence "Bring me a slab!" We want to say that he really MEANS "Bring me a slab!" But how does "Slab!" come to MEAN "Bring me a slab!"? What is it that makes "Bring me a slab!" a more "complete" sentence than "Slab!"? That it is four words instead of one? But does these extra words float before us in our mind somehow when we say "Slab!" and "mean" "Bring me a slab!"?

Suppose there was a foreigner who did not know our language very well but became accustomed to this command. He actually believed that this sound-sequence "Bring-me-a-slab" was one long word, which might for example correspond to the word for "building stone" in his language. We might think that he is leaving out something, namely the conception of the other words in the sentence. But do WE actually have those words conscious in our mind when we say "Slab!" and "mean" "Bring me a slab!"?

All of this confusion ultimately rests on our tendency to think that there must be some inner structure in our mind which is the "real" meaning of the form of the sentence that we "mean." We are thinking of "understanding" or "comprehension" as some kind of state that exists in the subject's mind. We want this to be what ultimately distinguishes our saying "Bring me a slab!" and "meaning" it as four words from the foreigner who says "Bring me a slab!" and "means" it as one word. But what actually distinguishes these is not what occurs in the subject's mind, since neither we nor the foreigner are likely to be focusing out attention on the grammar of our sentence at this moment. in the first place.

What ultimatley makes us "mean" the sentence in a different way is the different forms of life that our using it occurs in. It is not anything that accompanies the words as they are said, nor even the grammatical form of the sentences they occur in.

When we view language merely through the lens of grammar, it appears that it functions in quite a limited way. All sentences in terms of grammar can basically be boiled down to being one of three forms:
1. assertions (i.e. "the sky is blue")
2. questions (i.e. "did you eat breakfast?")
3. commands (i.e. "put your socks on")
But we know that this is a completely impoverished view of all the ways that we actually USE language in the endless variety of language-games that there are. Wittgenstein gives us a sense of this:

We get lost when we try to look for the "essence" of certain parts of our language and ask something like "What is a question?" or "What is a description?" That is because when do so, we ignore the fact that even if we limit ourselves only to questioning and describing, there are an innumerable number of ways to question or to describe in all our different language-games.

"Describing" a body's position by means of its coordinates is very different from "describing" a facial expression or mood. A police detective asking "questions" to a murder suspect is very different from a teacher asking "questions" to a student. Calling all these things "descriptions" or "questions" does not make all these distinct language-games any more similar to each other.

Marie McGinn points out that not only does this "craving for generality" get us chasing after illusory things like "the essence of a question," but it actually makes us ignore the real complexities and differences that are found in our everyday use of language. [3] These differences are not just incidental to language. They are the very foundation of language. The entire "essence" of our language is located in the fact that language emerges inside of language-games and that it allows us to understand and partake in these language-games.

In the Augustinian conception, language essentially amounts to attaching names to objects. But there is not just one way even refer to objects with names. Think about all the different kinds of names we use for things:

"table" <--- used to refer to all objects of its kind
"red" <--- used to refer to the color of a variety of things
"square" <--- used to refer to the shape of a variety of things
"that man at the bar" <--- used to refer to one specific human in front of us
"Bertrand Russell" <--- used to refer to one specific human who is no longer in front of anyone because he has been dead for deacades

These "names" are not used in any ONE single, uniform way. In fact, the variety of different techniques we use to refer to things around us might make wonder how we manage to ostensively define things at all in the first place. Wittgenstein gives words to these skeptical concerns:

The point here is not that we can never ostensively define things. We clearly can. We succeed in doing so with remarkable regularity, even including abstract things like colors or numbers. But the connection here appears mysterious. We could explain it more concretely if we said "this NUMBER is callde two" or "this COLOR is called red." As Wittgenstein puts it, this is essentially showing us the "post" at which we "station" the word grammatically. But if we can't explain the comparatively more concrete examples of "two" and "red," do we really expect to be able to explain something more abstract like the concept of a "number" or "color" in general? However, we only make this picture of defining and understanding a word seem strange because we isolate it from the form of life in which it occurs.

If someone was learning to play the game of chess, under what circumstances would I be able to point to a particular piece, say "This is the king," and have them correctly understand that I don't mean this color, this shape, this texture, this piece of wood alone and not others that look or function like it, and so on? The answer is that he must already have some mastery of chess. He might have been taught about the game out of a manual with diagrams. He might have just sat and observed enough chess games to have a general idea of how it is played. Whatever it is, he will have more of a familiarity with chess than a tribesman who has just walked out of a jungle and has never seen a chess piece before in his life.

To generalize, it is not so much what happens in the mind of a subject when he hears a word defined to him that makes a difference whether he succeeds in understanding it or not, but the whole surrounding circumstances such as what condition he was tuahgt in and how he goes on to use the term afterwards. Augustine describes a child coming to learn its first language almost like a foreigner in a new conutry who knows how to reason and to think and to express himself but just does not have the words to do it. He ignores the fact that a lot has to already be in place before we can even come to see someone pointing at something and uttering sounds as a "definition" of any kind. In short, there must already be a language-game that the child is participating in.

We might think that a child does not have to be fully aware of its surrounding language-game, but can simply "guess" what is being pointed at. Of course, there's no physical sign whether someone is pointing at a piece of paper or pointing at its color or shape or number. So the child will "guess" by focusing on the color over the shape, for example. But even something as simple as "focusing on the color" can mean so many different things that speak to language-games of their own:

We find no physical difference between "pointing at" or "meaning" the color rather than the shape or the number. But even though there is no physical process to distinguish them, the difference between the two is clearly still palpable. Therefore we're tempted that there must be a different mental or "spiritual" process in its place. But even if it was the case that a particular mental state did occur every time we pointed at or meant the color instead of the number, this would not be enough to ground the distinctions we're really after.

The real distinction has to lie in what happens before and after the pointing, not in the exact moment of pointing. It has to do with the entire form of life around us, not just our private consciousness.The truth is that everything we need to explain this differences already lies before us in the open in our language. We forget this and start to invent strange metaphysical pictures where we don't need them.

4.b. §38-64: Naming Objects

Wittgenstein now lingers on the concept of "naming" and the way names are "attached" to objects. Here he returns to some of the problems of reference that motivated Bertrand Russell and his own Tractatus Logico-Philosophicus. In his motivation to construct a fully analyzed logical language, Bertrand Russell began to explain that the names we use to refer to things are rarely definite or certain. The only truly logically proper names could be those which have no function in a proposition other than directly indicating an object. Russell claims that names like "Socrates" and even nouns like "chair" have to really be definite descriptions in disguise, as their meaning cannot be established independently of all other words.

        "Socrates"                
        ─────┬─────               
             │                    
      ┌──────┴─────┐              
      │            │              
   Plato's       teacher          
  ──┬──────      ────┬───         
    │                │            
    │                │            
a Greek          a person who     
─ ─────          ─ ────── ───     
philosopher      instructs a pupil
───────────      ────────  ─ ─────
                                  
           [...]

   "this"	   "that"

For Russell, the only truly logically proper names are "indexicals," meaning words like "this" and "that." These merely pick out the most primitive bits of sense-data that we are directly acquainted with. Therefore, they truly do function like names as they simply refer to what is in front of us with no connection to any other things.

There appears to be some similarity between the use of names and indexicals. In many cases we can use them interchangeably. I can say "Mount Everest is tall" and I can also say "This is tall" if I am pointing at Mount Everest. But we take this a step too far if we assume that this means that "this" is a name. The language-games of "this" and a name are actually very different. For example, we we can only use "this" in the presence of the object that it indicates while we can use names in the absence of what they indicate. "This" has no meaning without ostensive indication, but names do have meaning apart from ostensive indications.

P.M.S. Hacker compares indexicals to the point of origin in a system of coordinates (i.e., the (0,0) values on a scatter plot). They fix an index by reference to which related terms can be used that are relative to the index, like "over there," "yesterday," "N miles away," and so on. Names do not function like this. [4]

Russell adheres to the idea that names can be indefinite and vague but "this" and "that" are definitve. Wittgenstein's own version of this concept of logical atomism (as put forward in the Tractatus) differs from Russell in many important ways, but it is also based on the idea of reducing complex names and signs into the most simple and indestructible objects possible that stand as they are independently of any other state of affairs. Wittgenstein is thus critiquing his own early thought as much as Russell here.

The argument runs that the "definiteness" and "certainty" of experiencing something like "red" for example ultimately happens at the level of our grammar. There are certain things that make no sense to doubt in our language-game, but also make no sense to be certain of. It looks at first glance like individual red things could exist or not in a contingent way but that the color "red" itself had to be eternal and indestructible. But when we say "red cannot be destroyed," this is actually a statement about the USE of red in our language. We mistake an understanding about its role in our language game for an understanding about some metaphysical status of "red."

If we say that "red" is unable to be defined, it means that it is unable to be defined ANALYTICALLY. We cannot describe it as a substance with some particular independent quality, but we can describe "red" in ways like saying it is the color of blood or the color complimentary to green or the color pink turns into when it gets darker and darker and so on, as P.M.S. Hacker says. [5] These are all perfectly functional "definitions" in the context of our ordinary language-games. When we explain what "red" is, we might point to something that we call "red." But this does not mean we are pointing out some phantom entity called "red," which is the "real" meaning of the term "red." Our pointing does not show us some entity to which we attach a name tag. It shows us how to USE the word.

4.c. §65-88: Family Resemblances

Up until now, Wittgenstein has "answered" the problems of philosophy by dissolving them and showing that the answers we seek are already found in our language-games. But an explicit definition has never been given for what a "language-game" is. He never explicitly says what is common to all language-games and what the essential definition of what a language-game is. In this way, he seems to continue to skirt around the issue that he raised in the Tractatus and pourd over for so much of his career: the general form of a proposition and language. He gives no hint as to what is essential to a language-game itself but only gives examples OF language-games.

However, this is actually exactly as things should proceed in philosophy. Since we use the term language-"game," we should think about the word "game" in the first place. Also note that the original German is "Spiel," which also basically means "play" and therefore has a somewhat broader use than the English "game."

We seem to be unable to find anything common to all "games," and yet the concept still makes sense to us. We find ourselves with an intuitive idea of where "game" applies and where it does not. Wittgenstein's radical move is to extend this to all concepts in general. The labor of the Tractatus to find the essential form of a proposition is now thoroughly rejected. The meaning of any concept is no longer to be thought of as an ideal, essential core of shared characteristics, but as an overlapping of similar characteristics whose edges may be blurry. What Wittgenstein wants to show us is that a concept having blurry edges by no means makes it meaningless or useless. He compares it to the way we speak of the members of a family resembling each other:

Concepts can expand, contract, and evolve over time. To use Hacker's example, the concept of "art" had to expand to include photography and film when they were invented. [6] The same goes for all concepts. When philosophers find words used in a vague manner that is difficult to unite in one way, they are tempted to "fill in the gaps" and "sharpen up" the edges by creating a clear definition. But this ignores the fact that it is precisely this vagueness and blurriness that gives these words their characteristic meaning in the first place!

One can "draw boundaries" around these words for some special purpose or another, but these boundaries are not necessary for the concepts to be of use to us. In Wittgenstein's example, I don't need to be able to give a complete definition of what a plant is for someone to know what I mean when I say "the ground was covered with plants."

An interlocutor may raise the objection that we come to know what a concept is by seeing something that all examples of the word have in common. The clearest example of this is a color like "ochre." We will look at various pictures with something of that color and come to understnad it as "ochre." We might also think of showing someone various different leaves until he has a sort of schematic mental picture of what a generic "leaf" looks like. But in this case, how is that to be understood as a "schema" in the first place and not as the shape of a particular leaf? How would the ochre we have in our mind be understood as "pure" ochre and not as some particular sample? This all comes back to how the samples are actually USED, and this brings us back to the openness of family-resemblances.

Wittgenstein makes it clear that a concept being "open," like concept families are, does not necessarily mean it is "vague" in every way. There can indeed be sharp boundaries even when the concept is open. In the case of numbers for example, cardinal numbers, rational numbers, real numbers, and so on are all clearly defined with any individual number providing an obvious yes-or-no example. And yet, the concept of a "number" in general is exceedingly open, as it has changed and expanded quite a lot as mathematics has developed.

Does the idea of a concept being in some places "open" and in some plases very rigidly bounded seem nonsensical? It should not. Keep a game like tennis in mind. There are no rules for how high or how hard one may throw the ball, and yet there are still other definite rules that keep it "bounded" all the same.

In fact, in one sense it seems absurd to even have an idea of a concept where we would have rules pre-ordained for every possible situation where this term could be used. Wittgenstein gives an example with the use of the word "chair:"

This should show that "understanding" a sign or concept by no means entails that there is no room for doubt in application of its rules. In fact, Wittgenstein claims that a fully ideal language with perfectly understood logic and no ambiguity would have to be a language held in a vacuum that cannot possibly exist in the real world. There is always a possibility for rules to be ambiguous and to be reinterpreted. He compares a rule to a signpost on the road. Where is it written whether we are to follow it in the direction it points or the opposite one? What if there were many signposts? Or marks on the ground? Maybe there is no doubt in how to follow them. Maybe there is. But this is now an empirical question. And the degree to which they are doubted or not is all based on the surrounding forms of life.

4.d. §89-137: Philosophy as Therapy

This section is a broad sweep consisting of a number of meta-level remarks on the overall method and goal of this text and of Wittgenstein's later philosophy in general. It might have made more sense to put this section at the beginning of the book, but it actually makes more sense after we have seen the method at work. Much of this is still based on a critique of the philosophical method of the Tractatus Logico-Philosophicus, particularly of the idea that logic states THE single essence of language. But Wittgenstein's change here is more radical than may have been clear up until now.

For him, a search for the logical essence of language has to be replcaed by a study of how we actually USE Our language in everyday existence. But this ultimately means that in a successful application of this technique, there will be no new "knowledge" and no "answers" to philosophical problems whatsoever. Most of this section consists of explaining this method over and over in a number of pithy aphorisms. And that method is a "grammatical" investigation.

This means something more borad than "grammar" in the sense of proper rules for constructing sentences, conjugating words, and so on. It means a look at how words are used and how language is practiced. Of course, this is inseparably connected to an understanding of the forms of life they emerge in.

We have to be reminded of what we already know about how we use words. When we try to set up some elaborate picture of a philosophical problem like "What is a thought?," we are setting up a theoretic model that suddenly becomes mysterious. But we know exactly how words like "thought" should be used in our everyday speech. And for Wittgenstein, THIS is what their real meaning ultimately is. The whole direction of his later philosophy is the reverse of how we normally consider philosophy to be practiced. He leads us away from theory and back towards the everyday:

What would the point be in doing something like this? What value does philosophy have if this is its goal? Well, it wouldn't be anything like the traditional picture of the goal of philosophy. It wouldn't be anything like the goal of science which is the accumulation and discovery of new knowledge. Wittgenstein claims the method is essentially "therapeutic" (in the medical sense of the word). Its goal is to relieve us of the burden of strange models of the world we start to construct, because most of these are really empty and don't make any connection with what they're meant to explain in the first place. They are usually distorted images we come to have on the basis of being misled by our language.

In an ideal application of this method, all philosophical problems should completely disappear. And then, in the words of the Tractatus, there is no question left, "and just this is the answer." Wittgenstein feels that this shift is so radical that it needs to be sketched out over and over again in a number of metaphors and aphorisms. Here is a selection of the ones I find most helpful and thought-provoking:

4.e. §138-242: Understanding and Rule-Following

To return to issues at hand with a greater idea of the method and goal of Wittgenstein's therapeutic method of philosophy, we can address what seems like a problem: How do we come to understand a word or concept if the meaning of it is found in its use? When we think of our process of "understanding," it seems that it happens in an instant. However, the use of a word has to happen over the course of time.

When I understand a word like "cube," does the whole range of possible uses somehow come before my mind in a flash? We are tempted to think so. We might think that there is a certain use that becomes "ordained" in our mind that directs and comepls all future use of this word. Of course, this makes no sense when we start to think about it. We can easily imagine a case where we, for example, have a schematic picture of a cube in our mind, but point to a triangular prism. This is unlikely, but what exactly is it that "compels" us to only apply the mental picture we have of a cube in one way?

What is really going on here is just that we only have one potential use of "cube" in mind. We misunderstand this lack of other applications coming to mind as the "image" in our mind somehow "compelling" us. But the whole idea of an image compelling us seems more nonsensical the more we think about it. The reason we have this mistaken idea is that we are tmepted to think of our "understanding" as a mental state, which is the source of our ability to use words correctly. Maybe we don't think that our understanding outright "forces" us to apply the word in a certain way, but we might think that there is nevertheless some kind of internal mental state that gives us the ability to use language correctly or not.

Let's look at an example: Imagine that we teach a pupil to write a series of numbers according to a certain rule, like the series of multiples of 5 (5, 10, 15, 20, 25, etc.). We can judge if he understands it bsed on his being able to consinue the series to a certain degree with a certain level of accuracy. We are now tempted to say that "understanding" is a sort of determinate state that lies within the pupil's mind and whose outer signs are his being able to coninue the series as we expect him to. Once again though, this just seems like another way of saying that there is a certain something in his mind "compelling" him to apply the concept in a certain way.

Maybe the formula to continue the series comes to his mind. Maybe it has to, in fact, for us to say that he has understood the series. But it is still only on the basis of him USING that formula in one way or another that means he has "understood." After all, it is quite possible for the formula to come to mind and for him still to not understand the series.

Wittgenstein's imaginary interlocutor is unconvinced. He claims that it makes no sense to have to watch what we do in order to know what we mean by a word. How could that work? I have to know what I mean by a word before I can apply it! But again, the only reason we get into this issue is that we are thinking of "understanding" as a mental state that is the source of our correct use of word. When we really look at how the concept of "understanding" is used in our language in an honest way, it just doesn't fit that picture. It does not function like a conscious mental state, nor does it even function like a kind of physiological mechanism of the brain or any kind of "unconscious" mental state.

To closely paraphrase Marie McGinn, when we think of conscious mental states like being dejected, being excited, being in pain, and so on, we can speak of them in terms of duration, interruption, continuity, and so on. We can also speak in terms of intensity and degree. But when we speak of "understanding" we speak more in terms of capacity and breadth. Sometimes we can date the beginning of our understanding. But to speak of it in the same terms of duration still seems inaccurate. We could say that "understanding" something is a "state," but we don't see the grammar bear it out as a MENTAL state. Nor does it make it any more accurate to our grammar if we talk of it being unconscious rather than conscious. [7]

Imagine the following scenario: Teacher A writes down a series of numbers and pupil B finds a rule for continuing the series and then states "Now I understand!" We are tempted here if nowhere else to say that the "understanding" is a state that appears all at once in a flash and then explains B's ability to continue the series. Let's say that A writes down "1, 5, 11, 19, 29" and B says that he knows how to go on. What actually goes on in that case? Maybe B found out that the formula was "x(n) = n^2 + n - 1" after testing various other formulae. Maybe he realizes that the differences between the numbers are 4, 6, 8, 10. Maybe he says "Yes, I know that seires" and continue without further thought. Maybe he says nothing and simply continues the series with a feeling that may be called "That's easy!" But none of these processes alone can be called "understanding," because we could imagine every single one of them occurring without the speaking having understood.

Understanding that, we come to think that "understanding" must be something that goes on inside the speaker and not just in his behavior. Of course, this makes it even harder to apprehend when he understands it or not. Now we may ask how it is proper to say we have understood. If it is not the formula on its own that allows me to know that I am correct, then what is? As usual, it is everything SURROUNDING us. It is the circumstances that we are in. What is it that motivates me to use certain words, gives those words their meaning, and allows others to respond to them? The entire form of life that we grow up in and are trained into.

Note that Wittgenstein is NOT arguing that all understanding is really nothing but behavior. That is an obviously absurd conclusion. He is only arguing that without the accompanying form of life where we are trained to respond and perform in certain ways, it will be an incomplete picture. It may indeed be that mental states may be an essential part of this understanding. But it is the fact that this formula occurs in my mind when I am continuing this series rather than, say, when I am doing my laundry that marks it as a part of my "understanding." Understanding is ultimately a pattern of behavior that exists within a whole web of circumstances. Removing it from these forms of life takes us away from its essence rather than moving us towards it.

Now Wittgenstein wishes to complicate the picture. Let's say that our pupil is writing out the series "+2" and he continues it in just the way we expect: "0, 2, 4, 6, 8, 10..." We have tested his understanding up to 1000. But now he starts continuing the series as "1000, 1004, 1008, 1012..." We say "You are supposed to go on in the same way!" He replies "But I DID go on in the same way!" For it might be that this pupil interpreted "+2" as we would understand "+2 up to 1000, +4 up to 2000, +6 up to 3000, etc." How could we argue with him? We could only point to past examples which are less than 1000. And he would agree with all of these. Certianly we can't appeal to the idea of the platonic "rule" itself compelling us in our mind.

In this case, Wittgenstein says that it would be as impossible to argue against his interpretation as it would to argue against someone who naturally interpreted a pointing hand by looking in the direction from fingertip to wrist rather than from wrist to fingertip. All I can say at this point is that I "meant" for him to continue with "1000, 1002, 1004, 1006..." Of course, I didn't have these exact numbrs in mind when I "meant" it. And if by chance I did, there were equally many others that I did not have in mind. And the pupil could show equal discrepancies for these. This case is chosen for the apparent "exactitude" and "certainty" that a mathematical series implies. But it can be generalized to things like the use of words as well.

How was I able to prescribe all possible uses of "+2" when I "meant" it? How am I able to prescribe all possible uses of the word "cube," for example, when I "understand" it? We only get confused and find these questions vexing because we think of our "meaning" or "knowing" the infinite numbers in the series or infinite amount of uses of the word "cube" as something that is somehow already present in the moment of meaning or understanding.

The fact is that, yes, we DO mean ALL instances of "+2" or "cube," but they do not need to come before our mind for this to be the case. How can this be? In just the same way that we can say "I want to play chess" without all the rules of chess coming before our mind in an instant. The connection between that statement and all the rules and possible outcomes of the game of chess is present not in our mind but in the surrounding form of life that we are acculturated into.

If we look at the rule and the possibility of following it in isolation, it looks as though there is an alarmingly skeptical conclusion: That any rule can be made to accord with any exercise of it. This seems to be the implication if there is nothing that "compels" a certain use that stands within the rule itself. For example, the pupil could interpret "+2" in the strange way that continues with "1000, 1004, 1008, 1012..." and we cannot argue against him. We could interpret "cube" in a way that we point at a triangular prism when we hear it. But this problem only occurs when we make the philosophically natural but in fact strange choice to isolate an interpretation of a rule from all the forms of life that accompany it. As he puts it:

Ultimately, what this proves is that there is a way of grasping a rule and saying whether it is correctly followed or not that does not involve any "interpretation" of it. The rule can indeed be interpreted in any way. But at some level, the way we respond to simply amounts to OBEDIENCE. How we've been trained to respond to a rule is something that emerges in the form of life we are brought up in.

In Wittgenstein's words, we are prone to think of the beginning of a series like "O, 2, 4, 6, 8, 10" as a visible section of "rails invisibly laid to infinity." That is, of the unlimited number of future applications of a rule. Now this begins to look like a strange picture. All the steps have somehow already been taken and pre-ordained for my future use. But this is just an imaginative metaphor that we mistake for being something contained in the rule or our understanding of the rule. All this image of "rails laid to infinity" really amounts to is a way that we make sense of the fact that we no longer CHOOSE how to apply this series, but apply it "unthinkingly" since no other way to apply it comes to mind. The idea of "rails laid to infinity" is merely what Wittgenstein calls a "mythological description of the use of a rule."

In truth, nothing in the rule ordains us or compels us to use it in a certain way, not even our interpretation. But there is very much a right and wrong way to use it regardless. And all of that exists in the language-game and form of life that the rule develops inside of.

4.f. §243-275: The Private Language Argument

The part of the Philosophical Investigations broadly spanning from §243-275 is known as the "private language argument" and is one of the most famous and consequential parts of the text. However, it is not as tightly structured of an "argument" as the name might imply, and there remains a great deal of disagreement about what exactly the implications of this argument are and even where its beginning and end boundaries in the text are. What is clear is that the argument seems to imply that it is impossible for there to be a sensical, functional language that is not somehow shared or public.

The above section made it clear that there must be some general similarity in human behavior in order for our language-games to function. That is the reason we can use words or concepts in only particular ways and not others. It is not something that is pre-ordained in the rules or words themselves that we understand in a flash, but something that exists at a social level that we grow into. Wittgenstein now takes the argument a step further. He now says that in the absence of such generally shared agreement in behavior, not only can there be no rules for how to use language or concepts, there can be no "language" or "concepts" at all.

Traditionally, we would assume that there are certain sensations that we come to know and understand through pure introspection alone. Pain is an example. It is thought that we could identify and recognize certain types of "pain" even if we had never been taught the concept of "pain" and words for pains in our language. For example, if there were a feral child who had grown up in the wild, never encountering other humans or learning language, it would still come to have a concept of pain based merely on introspecting on its private sensations. We essentially define it ostensively by "pointing" and saying "This is pain." But here we point at something with our internal attention instead of pointing at something physical out in the world with our finger. Wittgenstein does not deny that such an introspection and awareness of internal states can exist. He denies that we can ever come to understand a concept or give meaning to a word on the basis of them.

A truly "private" language would be one that was completely subjective and phenomenological. It would only create words and uses for them based on private internal states. It would therefore be in principle understandable to only one person. If we believe that our understanding and naming of sensations like "pain" are ultimately derived from private internal experiences, then such a language would in principle be possible. But Wittgenstein days that since even concepts like "pain" are understood via public forms of life, such a private language is impossible. [8] [9]

When we look at how we actually learn and come to use terms like "pain," we notice that there is never a step of consciously focusing inward on some internal state. This is important. Wittgenstein does not say anything as reductive as the idea that all pain really amounts to pain-behavior like crying. Nor can we say that the "real" meaning of pain is just physiological processes like nerve-endings sending neurotransmitters to the brain. That would make it hard to explain why the concept of "pain" clearly had meaning before that specific bit of human physiology was discovered. An internal awareness can certainly be important, but it is not this particular internal mental state that is the "meaning" of a word like "pain." There is no need for some internal ostensive pointing to make a coherent connection between the feeling of pain and the words "I'm in pain," just as there is no need for it to connect pain to the act of crying.

The interlocutor may still be unsatisfied at this point. He might say that people can experience my crying but not my pain. There seems to be some distinction here between pain, which is private, and crying, which is public. However, this picture starts to dissolve when we look at it more closely:

There is indeed a distinction to be drawn here between what is private and public. But we get misled when we think of pain as being some "object" that is only accessible for the "subject" who feels it. The difference is actually all found in the grammar and not in the actual internal states. We can say "I'm in pain" as a way to express pain, just like when we cry. But we cannot express pain by saying "She's in pain." We cannot "learn" of our own pain, but others can. I cannot "doubt" that I am in pain, but others can. This is where the distinction lies. That distinction we are trying to make by saying that my pain is something private is not something I learn through direct empirical perception of some internal state. It is something I learn and come to express through the grammar we use to refer to sensation-concepts versus behavior-concepts.

The interlocutor may say that someone cannot have the same pains as him. The fact is, however, that in our everyday speech we often can say that people have the "same" pain as us. We can say that, for example, about a group of patients in a doctor's waiting room who are all experiencing the "same" pain in their stomach. However, the interlocutor feels that we are here missing the point. They are experiencing COMPARABLE pains, but not IDENTICAL ones. He claims that another person can't have "my" pains. But how do I determine what are "my" pains? How do I lay claim to them? By just focusing very hard on them?

The truth of the matter is that these confusions are all based on a misapplication and grammatical misunderstanding of the word "pain." This conception treats "pain" like a physical object like a chair that we can count the number of and identify as being the same or distinct from other objects, possibly being owned by different people, or located in different places and so on. But in our actual ordinary speech we do not count pains, nor do we identify or lay claim to them.

The interlocutor might be tempted to say that another person cannot have "my" pain because the pain is located in my arm or my stomach or my forehead and not his. Therefore he cannot have the same pain as me. But this makes just as much sense as saying that someone cannot have the same hair color as me because the color is located in the hairs on my head and not his. Wy does this strike us as odd? Because we do not use color in a way that involves counting and laying claim to it in our language-game. But we do not do that with pain either! We simply get confused and start to think that what "pain" means consists in pointing inward at some determinate mental state. [10] [11]

The interlocutor might assume here that a private language could never be communicated to others, but that it would still be possible to associate a name with a sensation in a totally private fashion. The problem is that even this most minimal action seems nonsensical. In the traditional Augustinian picture of language it might make sense to attach names to objects that come to us ready to fit into blank spaces that are already there in our grammar. But the problem here is that naming presupposes a language-game (a grammar of public use).

We never ascribe a name to a color, for example, just in order to call that color by its name. We come to name things like colors in order to distinguish a red apple from a green one, to say that the sky is clear and beautiful, or to say that the color on our painters' palette should be lighter, and so on. Naming has no use apart from these kinds of purposes. And since meaning comes down to use, the idea of a language without a public use becomes totally meaningless.

While it strikes us as strange to say, the above ritual of marking "S" for a sensation in our diary does not succeed in giving a meaning to that sign, nor in providing a rule for how it is to be used further. The most important point here is that there is no real way to check whether we have used the sign correctly or incorrectly. For anything to have meaning, it has to have situations where it is used correctly or incorrectly. This doesn't even have anything to do with having clear-cut boundaries of use or not. For example, it is very clear-cut whether a number is rational or irrational. It is not so clear-cut whether a painting counts as beautiful or not. But in both cases we establish the use of these words, vague though the latter may be, by public language-games. We understand their use by the similarity or difference between how we use them and how others do. But there is simply nothing like this to check our use against in the case of our sign "S."

If we were to go on using "S" correctly according to our private language-game, we would need to know that we are using it with reference to the same sensation. And how do we establish that the sensation we have now when we write "S" is the same as the one we had yesterday when we wrote "S?" There's no way to say what the difference between having sensation "S" is and only SEEMING to have sensation "S" would be. The moment we try to tie it to something measurable like, for example, our blood pressure rising when we have that sensation, we have turned it into a public language. This is because we have some means of verifying whether we have guessed correctly or not (i.e. whether our blood pressure really did rise or if it only seemed like it did).

In fact, on what authority do we even call "S" a SENSATION? We only call things "sensations" on the basis of them playing the same role as things like pain in our public language. But we lack any grammar of how to use "S." The truth is that when we have no clear criterion for how "S" is to be used, there is no way to justify the connection of "S" with any concept, not just a sensation. The whole idea of a private language dissolves before us the more we think about it.

But the real upshot of this is that all of this critique applies not only to some ineffable, undefinable sensation like "S," but even to our ordinary concepts like "pain." Even with "pain," the internal awareness of some definite state of mind actually plays no role at all in coming to understand what pain is. In short: a definition that is private is not a definition at all. Even if someone looks inward and says "pain" to himself, what role does that actually play in knowing if he understands the concept or not?

Wittgenstein says that asking we why cannot define a word privately is like asking why my right hand can't give my left hand money. It is simply not in accordance with the grammar of "give," just like looking inward and saying "pain" is not in accordance with the grammar of a definition. In any case, he makes it clear that it plays absolutely no role in the word actually having a meaning. He sums it up in a memorable analogy:

4.g. §276-315: The Beetle in the Box

The private language argument doesn't exactly have a clear end point. It starts slowly morphing into general reflections on the difference between "inner" and "outer" states of affairs, which are of course intimately connected to its consequences. Marie McGinn identifies this shift as occurring in §281, where we are again warned not to fall into the obviously false view that there is nothing to "pain" but "pain-behavior." It is obvious that one can indeed have pain and show no outer signs of it, just as one can hear something without hearing-behavior like turning the head in the direction of the sound. What Wittgensteins says is that these concepts of pain, hearing, consciousness, and so on are bound up with human forms of life and are incomprehensible outside of them:

The interlocutor here gives the example of children playing make-believe and imagining that their dolls are in pain to show that the concept of pain is in some sense detachable from humans. Even if we know that we're only pretending that the dolls are in pain, we're able to coherently talk about the concept of pain in something inanimate. Wittgenstein says that this objection fails because it is ultimately what might be called a secondary use of the concept of pain. That is, it still falls within the role of the wider human-oriented language-game. It is only possible to pretend that dolls are experiencing pain when we pretend that they are like humans. Ask yourself: Could we imagine that ONLY dolls experienced pain? Could "pain" have a meaning if it ONLY existed in our pretend game with dolls? That seems doubtful.

This whole idea is mistaken because it essentially treats pain as a kind of private object that we just happen to discover inside of ourselves and go on to hypothesize also exists inside of others. It turns pain into a kind of phenomenal "this." But this picture becomes strange as it essentially severs the idea of pain from the body. It depicts the pain as a psychological, phenomenal thing and the body as a physical, corporeal thing. The idea is then that the two somehow get linked in a way that remains mysterious.

The whole absurd view of pain as being a private, internal object comes from the tendency to think of the one who "has" pain now as a body but as a self or a soul. If we take this view seriously, it would make no more sense to ascribe pain to the body than it would to ascribe it to a stone or even to a number. The end result of this view is an extreme Cartesian dualism which results in the nonsensical idea of pain being capable of continuing apart from the body.

This idea is absurd. But it is a misapplication of a distinction which is indeed very real. But that distinction ultimately exists not between our "soul" and our "body," but between two different language-games. Namely, we use two very different language-games when talking about living objects and when talking about inanimate ones. We say that "I have a pain" or "He feels pain," not "My stomach has a pain" or "His ear feels pain." We are tempted by our grammar to assign the "I" or "he" in these sentences to something like a soul. And then we get into a hairy situation where we imagine the idea of pain being separate from the body. In truth, all the distinctions we need are already there in our grammar. We speak differently about living things and non-living things because of the different forms of life we exist with them in. And these distinct language-games are the bedrock of our explanation. [12] [13]

We might be hesitant to give up on the idea that there really is some form of inner identification going on that guides my use of the word "pain." It seems that we need to be able to "justify" our use of the word. But, as Wittgenstein says, to use a word without justification is not to use it wrongly. In fact, the more comfortable and competent we are at using language, the less likely we are to look for a justification for our use of it. We just use the words unreflectively to express what we feel. If we call this a "description" of our sensation, so be it. But it is clearly something very different from "describing" what the inside of a room looks like, for example. We should not let the similar wording obscure the different language-games being played in those two cases.

We might think that it is impossible to know if other people really experience what I do by "pain," and thus wonder how communication and mutual intelligibility is possible with this view. Wittgenstein responds to this with the most famous metaphor in the text. It is very self-explanatory:

We try to ground the distinction between pain and pain-behavior by construing pain as a private object and pain-behavior as a public expression of it. But if we choose to use this model, the "object" itself drops out as irrelevant. All the distinctions we are looking for between internal and external only exist at the level of our language-games.

At this point, Wittgenstein says that someone who is still obsessed with finding this private "something" is like someone who sees a picture of a pot with steam coming out of it and fells that we cannot understand the picture until we see what is boiling inside the pot. This is asburd. We understand that it is a picture of a boiling pot based on the significance of the juxtaposition of steam with a pot, not on our knowing anything about what is inside of it. It is very difficult to get us to be satisfied with the distinction between inner and outer and even of psychology versus behavior as existing at the level of our grammar, but ultimately that is where they exist most fully. We simply set up elaborate mythologies around our grammar, and this is where we get into philosophical confusion.

4.h. §316-362: Thinking and Talking to Oneself

From this point on, the Philosophical Investigations appears spottier and receives far less attention in terms of commentary. It becomes harder to sketch out explcit "arguments," to the degree that we can meaningfully speak of them in this text at all. The text is more fragmentary and indistinct. And yet, there is quite a lot here. At this point we are still only about halfway through the text. I will do my best to trace the arguments here, though it becomes more difficult. I follow Gordon P. Baker and P.M.S. Hacker's commentary quite closely and will do my best to cite them in an adequate way.

Wittgenstein now drifts from talking about sensations like "pain" to the phenomenon of "thought" as we conceptualize it. This is a natural transition, as we tend to have similarly Cartesian views of a kind of immaterial "self" or "soul" inside of us which is what does the thinking, just as it is supposedly what has the pain. But this way of thinking gets us off on the wrong foot:

We know that since the meaning of a word consists in its use, whenever we "point" at something, we are not giving its meaning but only clarifying it. We are showing what cannot be said. When we point and say "That is red," whatever we are pointing AT is not itself the "meaning" of "red." It is impossible to point at a meaning. We cannot point at any "thing" called "pain, but the concept of "pain" still makes sense.

There is some overlap here with the idea of "thinking." We cannot know WHAT we think by merely observing what happens in our minds. "Thinking" is not a report of a goings-on or an experience of something. Thought seems myserious and strange only when we start doing philosophy. But nothing about "thinking" seems strange until we start philosophizing or "thinking about thinking." We say that we pick up an umbrella because we think that it will rain. If you ask someone what they are thinking, they will answer as easily as if you asked what they are doing.

We only get confused when we start to believe that our thought is a kind of animating force that gives life to otherwise meaningless signs like words and sounds. We start to believe that "thought" is a kind of inner process that occurs in our head, a kind of ethereal light in our brain. However, in our daily life we can perfectly well use the verb "to think." We only get confused by the similar form of that verb to ones like "to speak" or "to talk," which are in fact used quite differently. So we make the mistake of believing that "to think" describes a sort of activity like speaking or writing or walking, but is one that goes on in the mind rather than in the mouth or the hands or the legs.

In fact, Wittgenstein wants to deny that "thinking" really describes an activity at all. Wittgenstein says that we think of "thought" as an incorporeal process that could be detached from speaking and other physical processes. He references a novella by Adelbert von Chamisso where the protagonist, Schlemihl, sells his shadow to the devil and now walks around without casting one. Of course, this is impossible to do in real life. Our shadow cannot be conceptualized apart from our body. We start thinking about "thought" in the same way, as if it were some strange phantasmagorical spirit-stuff going on inside of our skull that is different from the gray matter it is instantiated in.

Of course, much of this is sustained by our figures of speech, such as "I can't figure out what's going on in his head," or visual representations, such as cartoons with thought bubbles to represent what people are thinking. It is in fact useful to think about this example. We don't need a thought bubble when there is a speech bubble which is in accordance with it. We don't need a thought bubble over a character playing tennis with their full attention to know that they are thinking something of the form "I need to hit the ball!" (this example of thought-bubbles is taken directly from P.M.S. Hacker). [14] What does this imply?

Wittgenstein's simple point is that there is no one activity called "thinking" that we can isolate from other forms of life. We call a lot of things "thinking" which are clearly quite distinct: speaking before thinking, thinking to ourselves in the imagination, thinking up a solution to a problem, something unconsciously appearing in our mind when we hear a certain word... We might want to say that what is common to all of these is simply that the mind is engaged and active. But even in just a philosophical context, we really do a lot of things that get lumped into the category of "thinking" such as reflecting, musing, deliberating, speculating, reasoning, inferring, etc.

We say that "thinking" is an activity that occurs primarily in the brain. This is a distorted picture, but it has good reasons. As Wittgenstein puts it, it is a prejudice but is not a STUPID prejudice (§340). It makes sense to think of our brain as the "seat" of thought as it is right in the middle of many of the most important sense-organs that our experience depends on, like our eyes which see, our nose which smells, our mouth which speaks and tastes, our two hands which touch and grab and write, and so on. But again, this purely physiological picture doesn't quite capture what we mean when we use the verb "to think" in all the ways we normally do.

When we say what someone is thinking, we almost never talk about what is actually happening in the brain, any more than we talk about what happens in the larynx, tongue, or palate when we talk about someone is saying, or what happens in the phalanxes and joints of the finers when we talk about what song someone is playing on the piano. Elaborating on Wittgenstein, P.M.S. Hacker illustrates the point with the example of a person "thinking" about rearranging furniture in a room. What would happen if that person got a tape measure out and started measuring? Is he still "tinking" about rearranging the furniture? We would be inclined to say yes, even though an image of rearranged furniture would not be present in his mind's eye at that moment. [15]

Our grammar misleads us when we say that something is "thinking" X, Y, or Z. We rarely talk about what goes on in his mind. We talk about what he's aiming towards, or the conclusion or opinion he puts forward and its consequences or what he will resolve to do from here on out. We say "I'm thinking about publishing a book" and know that this is not a statement about some uninterrupted mental process but a statement of intent. When we say that "I think that slavery is wrong," we know that this is an expression of ethical condemnation towards another. When we say "I think you meant to say it like this," we know that we're teaching and guiding someone in the use of their words to avoid a further misunderstanding. A mere description of an internal process cannot have these sorts of consequences in our language-game.

While we cannot say that "thinking" is merely a description of what goes on in our minds, Wittgenstein is not saying that what goes on in the brain has nothing at all to do with thinking. That would be absurd. Clearly, when we "think," there are things that go on in our mind. But when we say "I am thinking about X" or "I think X," we are not describing the hodge-podge of images, words, phrases and so on that appear in our minds as this "X."

This takes care of some other pseudo-problems that might emerge when we have a faulty image of "thought." For example, it might seem as though we can "think" much faster than we speak, as we often speak of understanding something, say a logical argument, "in a flash" when it takes quite a bit longer to write out. We're tempted to think of this as the full picture of what we afterwards express in writing somehow happening in fast-forward in our mind. Wittgenstein says that we should actually understand it more like when we take a few short notes that sum up the core of an entire complex lecture.

In short, Wittgenstein denies that "thinking" describes a kind of internal, mechanical process that solely goes on in the head. In fact, he is skeptical of the idea that it describes a "process" at all. A more correct analog would not be a physical process like a heartbeat but the description of it as being something like a "frenzied" or "regular" heartbeat.

The question now arises: How does "thought" connect to language? We are tempted to think that linguistic signs like words and sounds are empty of life and life breathed into them by our minds. Wittgenstein disagrees. he thinks that what we call "thought" is inseparable from a number of activities in our life and that language is one of them. What gives linguistic signs their lives are, as usual, their USE in language-games.

We have the mistaken picture that we believe that we think "in" language. Do words appear in our head? Yes, we do talk to ourselves in our heads and quite often. But Wittgenstein argues here that we make an illegitimate leap when we say, because some things in our head take the form of an interior "saying," that ALL mental activity must have this model.

We might feel that our interior talking is essentially the same as our exterior talking but with the direction reversed. We know, for example, that when we attempt to speak in a language we are not very used to, we often pause to "translate" our words from our first language and sometimes get awkward phrasing as a result. This provides the allusion that we not only think "in" language, but "in" our native language until we become able to think "in" a foreign language. But this is a mistake. It is wrong to say that we think "in" anything, be it languages or images or whatever else. We can talk to ourselves in our imagination, but that is as different from real life talking as doing an experiment in the imagination is from doing an experiment in the lab, to quote P.M.S. Hacker. [16]

We use the phrase "talking to oneself" about someone if they can OTHERWISE talk normally. This is why we wonder if some internal talking exists in humans, but not in things that emit speech but are not capable of communication (participation in language-games) like parrots or gramophones. The chair example is revealing. We can only imagine a chair "talking to itself" if we start to imagine its back as a kind of head and so on. That is, we imagine it to resemble a human. For it is only in the context of a public language-game that privately "talking to ourself" makes sense.

4.i. §363-397: Mental Images and Imagination

The concept of "talking to ourself" is not learned or defined by the identification of some internal process. It is taught based on the language-game around us. We can see that our internal speech is still tied to our outer language-games insofar as we can say anything about it. The same even applies to what we call "imagination." Of course, from the beginning we are tempted to have a limited understanding of what it means to "imagine" something, given the etymological connection to the word "image."

The temptation is to define "imagination" as a mere conjuring act of placing images before the mind. We get misled because we begin to think of these private mental images as essentially the same as our public images. We think of our relation to them as essentially the same as our vision, but, as it were, happening on the inside rather than the outside. All of this totally obscures how the concept of "imagination," the concept of "imagining," and even the concept of internal visualization occurs within the language-game.

First things first, it should be obvious that it is an oversimplification to say that all "imagining" consists in apprehending mental images. It's true that we often speak of the "mind's eye," but we never speak of the "mind's ear" or "mind's nose," even though we can surely imagine sounds and smells as well. Nor do we need an image in our mind to "imagine" something. We can imagine someone being excited or disappointed to hear a particular piece of news without having a mental image of them in our head. To use one of P.M.S. Hacker's examples, we say that Christopher Columbus "imagined" that he had landed in the Indies without considering him to have necessarily had any mental image of the Indies appearing in his mind. [17]

We know that people certainly report being able to imagine mental images, although it is true that this ability exists on a spectrum with some people totally lacking it (aphantasia). But what gets us into trouble is when we make the leap and assume that mental images are essentially the same as real images or that our mental "vision" essentially works as an analog for our physical vision. The grammar between the two is actually incredibly distinct. This is clear in some ways that are so obvious that it seems almost silly to dwell on them, such as others not being able to "see" the same thing that we do when we see our mental images.

It is a mistake even to think that the "grammar" of mental images is more like each of use having a private peephole inside of our heads that only we have access to. A lot of questions that we ask about someone physically seeing something make no sense when we try to translate them to his mental vision. We cannot ask questions like "Are you sure you saw it correctly?" or "Can you take another look?" We want to talk about images in our mind as if they are things we can observe and empirically discover. But that is simply not how they function.

Just think about under what circumstances we would say that someone sees the Taj Mahal versus saying he IMAGINES the Taj Mahal. Under what different language-games do we tell someone to "look" and tell someone to "imagine?" We could ask if what someone was looking at was "really" the Taj Mahal based on whether or not it had certain features (color, size, etc.) But we could never doubt that what someone was IMAGINING was really the Taj Mahal, even if he made some mistakes in terms of the features being correct. It is not as though these errors suddenly make it no longer the Taj Mahal that he is imagining. When I say "I imagined it would be colder today," I am not seeing a picture before me and then dispassionatley describing it.

The reason that we call what appear in our minds "images" and think of them in such terms ultimately comes down to the language-games we use to discuss them. Wittgenstein here seems agnostic about to what degree coherent entities called "mental images" can be said to exist on the level of neuroscience. But he is clear that we get distorted in thinking about them. If they do exist, they would exist in the form of something like synapses between neurons. We get distorted when we imagine our skull as a kind of empty theater and the mental images as a kind of light show being projected inside of it.

Mental images get their coherence as a part of the broader language-game called "imagination." And what we can say about them ultimately has to amount to their grammatical USE. He gives some aphorisms here that apply generally to this conclusion and to the related ones of the last few sections:

Also note that when we speak of "imagination" in our daily lives, we almost never use it in a way that is concerned with what goes on in the mind alone. We are interested in imagination that serves some purpose. The moment that imagination turns into something else, like creativity, it starts to translate into the external world, into writing stories, painting pictures, and so on. "Pure" imagination seems to do almost nothing and therefore to be outside of the bounds of any meaningful role it could play in our language-game:

Pretty much all forms of imagination that have genuine consequences and thus can play a genuine role in our language-game do not simply stay in the mind. They must somehow be consequential in a public language-game. They must require us to behave or act in a certain way:

4.j. §398-427: The Self and Consciousness

At this point, Wittgenstein moves to some discussions about consciousness generally and implications for what is traditionally given the label "I," the soul, or the subject. Some of the most difficult parts of the Tractatus Logico-Philosophicus address this by arguing that psychical facts can be expressed in a coherent logical fashion without ever entertaining the idea of a "subject" to have them. The "ego" or "subject" completely drops out of the logic of language as something that can be meaningfully spoken of. As he said in that book, if I reported on everything I could encoutner and know in the world, if I wrote a book called "The World as I Found It," the self would be the one thing that I could not report on. I would isolate all the parts and processes of my body and the thing cald a "self" would shrink to an extensionless point with nothing to be meaningfully said about it.

The Philosophical Investigations echoes this theme. As might be predictable at this point, Wittgenstein wants to show that questions of the "self" or "ego" ultimately rest on a confusion of grammar and that the idea of "consciousness" exists entirely at the level of our language-games.

To begin to understand these thoughts, we can ask the question: What does it mean when we take someone's facial expression as "friendly?" Certainly we are only seeing certain physical features of the face. How do we see the "friendliness" in the impression? This seems to only make sense as long as we can have some access to a private mental object that is ours and ours alone. Wittgenstein enters into this discussion with some extremely difficult passages about a "visual room" that we imagine:

These passages are hard to parse, but the position that Wittgenstein responds to seems to be something like this: When there is a physical room, then we must have some private impression of it as well, a "visual" room. I must be able to describe this "visual room" to another, but I alone can have access to it. If I'm looking at the same object as someone who is colorblind, we will clearly be having a very different visual impression, even if what we are looking at identical. We clearly understand what someone means when they talk about their visual impression of something. But what is this "visual impression" if not something I can point to, look at, or so on?

The concept of a visual impression makes no sense when we try to treat it like an object. Our typical language-game of how we treat an object is clearly different from how we treat the visual impression. Again, the reasons here are mind-numblingly obvious: We cannot pick it up, others cannot access it, it occupies no physical space, etc. So what sense does it make to say that it "belongs" to anyone, much less to us?

In truth, the visual impression is not anything we can talk about as something that a "subject" relates to. What we call our visual impression is ultimately something that exists at the level of our language. When I say that my "visual impression" (my visual room in this case) is X, Y, or Z, this is really just a way of saying that "I see" X, Y, or Z. What seems like a discovery of some private mental object is really just a discovery of a way of speaking about our experience in our language-game. When I say that I have a visual impression, I do nothing more than express HOW something sruck me.

Now that we realize that we cannot give some "ownership" to the sense-impression, we might be tempted to say that the "self" is nothing but a bundle of perceptions, or that it is a logical construction made out of sense-data, or that it is a limit of the world but not something inside of it. But really these pictures only occur when we drift far away from our everyday language, just as much as the idea of the self as a Cartesian "soul" does. When we say "I have such-and-such a mental picture," we are not putting forward some metaphysical picture. We are not postulation the idea of an incorporeal entity in our minds. We are merely expressing ourselves. This way of talking does not signify any of the above metaphysical pictures.

At this point, Wittgenstein resurrects the point from the Tractatus that the arguments between "realists," "idealists," and "solipsists" do not really change the expression of anything consequential. But the Philosophical Investigations identifies the locus of their struggle not at the level of what facts can be put into words, but purely a the level of our forms of expression. When an idealist tells us that material objects are really only collections of ideas, does it follow that he thinks that we will not get burned if we put our hand in a fire? Of course not.

It could be possible to change our language to accommodate one philosophical "position" over another. We could replace all forms of "I think that..." with "The thought is now occurring that..." But this would actually change nothing about what is actually expressed. For these "positions" amount to little more than modes of expression. Wittgenstein claims that saying "I think that..." does not put forward the theory that "we" are a solitary subject any more than raising our hand in a crowd does. Because saying a sentence that starts with "I" is very much of the same nature as putting your hand up: A learned behavior to direct attention in our form of life. [18] [19]

Wittgenstein now moves to the question of "consciousness" generally rather than the individual "ego" or subject. Traditionally we think of consciousness as a kind of light that shines over gray matter that is dark and unthinking. We think of it as a kind of ethereal, airy stuff that permeates the dead matter around it. But this image is completely misleading.

There's nothing wrong with saying that consciousness emerges in increasingly complex organisms. There's a good reason that we consider human beings "conscious" but not trees or amoebas. But when we think of "consciousness" as a kind of strange, non-physical substance, it becomes hard to understand how something non-physical could emrege out of a physical substance. No matter how complex it is, a brain is just as much a physical, material thing as a stone.

The truth is that we never find this distinction between "conscious" and "unconscious" matter at a purely physiologial level. No matter of processes in the brain can bridge that gap. That is because "consciousness" exists at the level of our GRAMMAR. When we talk about "consciousness," we are talking about how we refer to living bodies versus how we refer to others that are not living. When we talk about human beings as being objects, we are clearly serving a special philosophical purpose in doing so, because that is not how we normally treat human beings and not how the grammar of our language-game refers to them. There is no deep mystery about how a living creature can be "conscious," just as there is no deep mystery when we say that we are "conscious" when we are awake and "unconscious" when we are asleep.

In some very artificial state of mind, I could imagine that everyone around me was a mere automaton. But this is not a state of mind I could stay in for very long. This is because everything in our language-game is built on the opposite assumption, and that is where the idea of "consciousness" actually exists. In Wittgenstein's words, it is my attitude that a man has a soul, not my opinion that he does.

There is an important distinction here between a similar earlier example where we imagined that everyone around us was in frightful pain but concealing it. That would also be quite hard to maintain in our daily life, as we would have to constantly act to sustain this very artificial state of mind. But we can actually imagine what it would look like to be in pain but pretend you are not. It is in some sense feasible.

But how could I say that the humans around me who were automata were only "pretending" to be conscious? What is it to "pretend" to be conscious? René Descartes once famously claimed that "I think, therefore I am," meaning that the one thing that could not be doubted was the fact that I am conscious, or perhaps more accurately that there is consciousness. But in fact, this undoubted axiom is just nonsensical. "Consciousness" only shows itself in the grammar of how we refer to certain kinds of beings. There is no way to "know" that we are conscious. Nor is there any way to "claim" that are conscious, to "seem" that we are conscious, or to "doubt" that we are conscious.

Consciousness exists only insofar as it becomes a way of talking about a human versus a stone, for example. But that is the only way it can have meaning. The only way we could say "I am conscious" in a way that plays any meaningful role in our language-game is if, for example, we say it when we wake up in a hospital after an anesthetic. [20] [21]

4.k. §428-465: Intentionality

For this section, it's important to distinguish two similar-sounding words that have different philosophical definitions:

"Intentionality" <--- a fancy term for our thoughts being "about" something or other

"Intention" <--- what we plan or project ourselves doing

This section is primarily about the former, "intentionality." But to make things even more confusing, Wittgenstein actually talks about "intention" not too long after talking about "intentionality!" But for now, we should focus on "intentionality." How are our thoughts ABOUT one thing and not another?

"Thought" is not to be conceived of as a mental state or process and especially not a the encountering of an "object" by a "subject." In the Tractatus, Wittgenstein wracked his brain to try to explain how a proposition could be false but still convey meaning. The thought went that it is hard to explain one state of affairs being true and I myself thinking of a false state of affairs in its place. How could I think of something that does not exist or is not the case? How could it exist in my mind if it did not exist in the world?

Wittgenstein attempted to answer this question by saying that the propositions we think of and the states of affairs in our world are "isomorphic." That is, they a structure. The arrangement of our thoughts somehow mirrors or resembles the real state of affairs, and it is on that basis that we can compare a false state of affairs with a real one that resembles it in some way but ont others. His analogy was of a ruler. We say how long something is by laying a ruler next to it and making use of the similarity between the object and ruler, even though the only thing the object and ruler have in common with each other is their length. So it is that the true or false proposition and real state of affairs only need to have their logical form in common.

The early Wittgenstein himself might as well be in the role of interlocutor at this point when he employs the ruler analogy again in order to show how we "give life" to the "dead" signs of words and sounds in our language:

Wittgenstein wants to say that if we view signs (words, sounds) as isolated, "dead" things, then they will be just as good to us as an alien script. It is a mistake to assume that signs themselves are dead, senseless things and that the important thing that gives signs their meaning is merely the "thought" accompanying them. It may be true that sometimes we can speak without thinking, such as when we recite the alphabet unconsciously. But it is a mistake to assume that it is thus the accompanying thoughts that alone give signs their meaning. It is the USE Of signs that imbues them with meaning.

Thoughtful, meaningful speech emerges in a form of life. The only thing that "animates" signs is the grammar of our language. The gap between language and reality is not something that exists anywhere but in our grammar. What makes a specific word into a sign? The same as what makes a piece of wood into a chess piece or a paper note into a unit of currency: The surrounding practices that it exists in the context of.

So how can I "think" or "mean" something that is not the case? How can I "wish" that someone will return home when that particular state of affairs is not something that exists anywhere in the world. How does my mind "link up" to it? How does it "reach out" into the future? This whole question is based on a misunderstanding about our language:

The mistaken idea all this emerges from is the idea that when we expect, wish, or fear something we are standing in relation to that something as a represented object that is contained in our thoughts. But our thoughts do not "contain" anything. How can a thought be "about" something that is not the case? This question cannot be answered, because it is misconstrued. Our thoughts are are not "about" anything. Our thoughts are not "representations" of some fact that is "out there" in the world or not, because facts are not "out there." Facts are not "found" or "contained" anywhere.

I can think of things that are not the case or have not happened yet. The problem only occurs because we're tempted to capture these things with a noun like "representation," "fact," "proposition," "event," and so on. No noun can fit in place as a "thing" which is the answer to a question like "What did you expect?," be it "image," "description," "idea," or whatever else. The only thing that can stand in that place is WHAT WE EXPECTED! And that just brings us back to the language-game, which is where all concepts of expecting, wishing, hoping, fearing, and so on get their real meaning. [22]

4.l. §466-490: The Senselessness of Skepticism

In the Tractatus, Wittgenstein took a very hardline stance against the idea of inductive reasoning. When I say "inductive reasoning," I mean reasoning based on the observance of past events. For example, if I put a coin in a vending machine and nothing comes out, I assume that the machine must be broken. But I only do that because something has come out every time I put a coin in the machine in the past. I assume it is not just chance that it happened not to work this time. But Wittgenstein claims that there is no logical reason to infer the future from the past. For him, there is only logical possibility and only logical necessity. No amount of consistency in past experience can be used to predict the future on a purely logical level.

For the early Wittgenstein, even something as simple as our belief in cause and effect breaks down when viewed from the perspective of pure logic. For him, things are simply ture or false. The number of times they occur or not and their proximity to other states of affairs is all contingent. I let go of a glas of water. Every time someone has done so in recorded history, it has fallen to the ground. What guarantees that it will not float up in the air now? Nothing. Only the number of times it has refused to float up in the air in the past, which is no logical guarantee at all, but simply a form of faith that the world will continue to be uniform.

What do we make of this idea in the Philosophical Investigations, now that the preeminence of logic is replaced with the shared behavior of language-games?

Wittgenstein opens this section by asking "What does man think for?," which might sem like a strange way to get into this topic. But what he attempts to do here is to show the distinctions in our language-game that we will blow over if we are too hasty. We are asking here for "explanations." We are asking for answers to questions of the form "why?"

But here, Wittgenstein wants us to pay attention to the difference in grammar when we ask for a "reason" versus when we ask for a "cause." We often use these terms interchangeably in philosophy, but rarely do in everyday life. When we ask why someone does something, we are not merely asking what CAUSED them to do it. We are asking for a "justification." We want to know someon's beliefs, desires, motivations, and so on. When we ask "Why did you do that?," we would merely laugh if the answer you gave had to do with the physical mechanisms of your brain and central nervous system.

There are many distinctions we can draw out of these differences in the language-games between reasons and causes. We can speak of a reason being justified or not. We can speak of there being a reason for doing X or Y regardless of whether X or Y actually occured. For example, you could say that someone had a good reason for walking out of a room after hearing a rude remark regardless of whether he actually walked out of the room or not. This makes no sense for a cause. It makes no sense to say that a cause is "justified" or not. It makes no sense to say that there was a cause for X but that X did not actually occur. It would not be a cause in that case, because its connection to the effect would be spurious.

Reasons, then, unlike causes, tend to refer to something that exists for humans to give justification for actions. In this case, the call for justification occurs because of this general skepticism about inductive reasoning. As Wittgenstein said, why do we make boilers according to certain calculations and not just leave their thickness to chance? Is it because more boilers exploded before we started making certain calculations? After all, it's not as though boilers NEVER explode according to our modern specification. What justification do we have that they will continue to explode with less regularity? What justification do have for the future resembling the past?

At this point, we could put forward a purely pragmatic defense: We have found it generally beneficial to build boilers like this in the past. But this just begs the question. This defense argues that we should believe that the future will resemble the past because it has generally been beneficial to believe that the future will resemble the past. But this just assumes that believing in the way we have will CONTINUE to be beneficial in the future. In other words, it still just assumes that the future will resemble the past! The whole point we are trying to get at is WHY we assume that the future will resemble the past. And we see nothing in nature that tells us that this will generally happen, any more than there is evidence that boilers made according to current specifications will continue to be safer than ones made before them.

What we can say is that this belief that the future will resemble the past and that nature is uniform is something that exists even in our most primitive reactions. Certainly we will not reach our hand into a fire twice, even if we have no logical certainty that the fire will always hurt us.

But can we call this a "belief?" Does a deer avoid a fire because it "believes" in the uniformity of nature? Does a peacock spread its feathers because it "believes" in the uniformity of nature? By that measure, do we avoid putting our hand on a hot plate because we "believe" in anything at all? Or is it our ACTION that lies at the bottom of this language-game of giving "reasons?" It seems that this is the case. It is the bedrock of shared common knowledge.

Why do we avoid touching a hot plate? A thousand reasons could be given:
"I don't want to hurt my hand"
"It's scalding hot"
"The hot plate is plugged in"
"Neurotransmitters will send unpleasant signals to the pain-receptors in my brain"
And so on. But this is all auxiliary and all happens in retrospect. None of these thoughts occur as a REASON for us not to put our hand on a hot plate. This is because so-called inductive "reasoning" is not really a method of "reasoning" at all. We will never find the rationality of arguing from past experience.

But does this mean that we can never really have "knowledge" based on inductive reasoning but only faith, as Wittgenstein argued in the Tractatus? Here, he seems to present a different conclusion. It is the case that we exist in a world which, until now, has exhibited enough regularity for the future to generally follow the past. It did not have to be this way. The universe could have been far more chaotic than it is. But the conditions of this world are the ones in which our language-game emerged.

This doesn't give us a "justification" for our inductive reasoning. But it does seem to render meaningless any objection to the predictions we make on the basis of that inductive reasoning. It is simply a part of our language-game at the most basic level. To say that we cannot predict the future based on evidence from the past is simply a denial of the very grammar of what counts as "prediction" or "justification" at this point. It is a denial to participate in our language-game. And whatever goes outside the bounds of our language-games must simply be called nonsense. [23]

4.m. §491-570: Nonsensical Sentences

Some things in our language seem to be patently impossible to give meaning to. It makes no sense to say we hear a color or to see a taste. A "five-sided square" is impossible to conceive of. The Tractatus Logico-Philosophicus took this idea very seriously and featured an entire ontology based on the idea of certain states of affairs being inherently nonsensical and thus excluded from any possible logic of language. It seems that we can distinguish between things that are logically possible from things that are logically impossible. That is, we can make sense of the proposition "This couch is purple" even if that couch is not pruple, but not the proposition "happiness is purple." It seems, then, that what is ruled out by our grammar cannot be merely contingent, but must be ordained in logic. This then seems like a case where our grammar cannot be arbitrary.

The later Wittgenstein rejects this idea as a symptom of our tendency to create elaborate metaphysical pictures. The Tractatus began with a highly abstracted ontology like this. In a picture like this, we think that everything that is possible is somehow "there" in a kind of shadow, latent manner. It is not as "there" as what is "actual," which is a subset of the possible, but somehow it is "out there" in the world for us intuit and access as a logical possibility in our mind. This is unlike what is logically impossible, which is just nowhere. We then begin to think that our grammar is somehow a mirror of all those possibilities that lurk in the darkness.

This view is distorted. The truth is that the grammar we use does not REFLECT the essential nature of things. When we lay down rules for the use of words, we do not look out anywhere and discover these empirically. We CREATE them and ENFORCE them within our language-game. In laying down rules for the use of words, we fix what it is to count as such-and-such a thing. This might change. For example, the idea of an "element" has changed with scientific discoveries. We no longer talk about simple things like fire and water as elements but about things like oxygen and helium. But that is not because we somehow were able to "link" our concept "up to" the facts. We never do that with our words. We only changed the rules of "element" in our language-game on the basis of convenience, or utility, or some other reason.

It makes no sense to say that what is "possible" is present in any way, even as a potential. There is nothing in reality that corresponds to something that is possible but not actual. In that case, what sense does it make to say that it is somehow "more real" than what is logically impossible? We have a fancy metaphor of everything possible being somehow "out there" in a platonic realm. But that doesn't reflect anything real.

Grammar never reflects any kind of logical order or pattern in nature. It does not reflect ANYTHING! It simply determines what makes sense, and that is all that it means to say that something is possible or impossible. When we say that something is "possible" or "impossible," what we really say is that it is either included in or excluded from our language-game. It tells us nothing about the states of affairs in the world at all.

Is the grammar of our language "arbitrary?" If that means that we don't justify it by looking at something external, then yes. The rules of language are like the rules of a game like chess. We do not discovre the rules of chess by a deep understanding of the "nature" or "essence" of pieces like the king, the knight, and the rook. The order is the exact opposite. The pieces are only established as a king, a knight, and a rook by the very rules themselves.

Perhaps we have a grammatical rule like "nothing can be both 1 meter long and 2 meters long at the same time." The form of this sentence looks at first glance like an empirical claim. It seems like a discovery we made about the logical structure of the world by observing the bounds of what it logically possible or impossible. It is actually a statement of RULES masquerading as an observation. It is telling us that we cannot USE this particular string of words to refer to something.

Note that the above "rule" also does so without referencing reality at all. In this sense, the rule may be called "arbitrary." We might feel that we are constrained and that no sense could ever come from us saying that something is both 1 meter long and 2 meters long at the same time. That feeling of restriction seems to feel like it reflects some "boundary" of logical form that exists in the world. But the truth is that nothing stops us from changing the rules of our language-game. We are only restricted in that there are no practices that exist for using this string of words. But that can change. To use Hacker's example, before the psychology of Freud, the idea of "unconscious motives" would have made as much sense as the idea of an "unconscious toothache." There was no place in the language-game for the idea of unconscious motives. But once the rules are changed, nothing about the old rules has any domain over them. But for all the reasons sketched out in the private language argument, this new use does not seem to be the kind of thing that one person can bring into existence all of a sudden. [24]

There are so many language-games that it is hard to find one phrase which will be nonsense in all of them. When something says something that is senseless like "I am cutting red into pieces!," we reply with "You cannot cut red into pieces!" This misleads us into thinking that we are finding some definite boundary of sense that is located "out there" in the world. But in fact, if a sentence is nonsense it makes no more sense to deny it than to affirm it, just as we could not make a chess move in response to someone putting a go-stone on the chess board. When we "deny" these claims, we are actually just clarifying the rules of our language-game. The whole problem of how we can "deny" a piece of nonsense need not occur.

4.n. §571-610: Psychological Experiences of Qualia

This section addresses some problems of psychology insofar as philosophy can elucidate them. It is pretty unfocused and vague, but there is value in it regardless. The comment that motivates most of this section is one made by psychologist Wolfgang Köhler. In 1929, Köhler argued that psychology was a young science and that its advancement would result in a more robustly theoretical treatment compared to what we have now:

This analogy of physics is taken directly from Köhler. He argued that in its infancy, physicists would have to observe the way things behaved in the world, like particular things falling to the ground. As the science matured, physicists were able to move into a more purely theoretical model which was no longer dependent on having raw observations in front of them. "Gravity" can now play a role in equations without us having to measure things out in the world as they fall to the gronud. None of our modern physics, which is highly dependent on precise calculation, would exist if leaps like this were not made. Köhler did not support immediately severing psychology from pure observation. In fact, he was highly critical of others in his day who lapsed into highly imaginative pictures on the basis of spurious evidence. But he did think that the direction of psychology would lead us towards something more theoretical as it matured allowing us to be less dependent on the raw observational data of individual people.

This parallel of psychology to physics is based on a misconception that psychology describes "psychical" processes in the way that physics describes physical processes. Much of the groundwork done to reject this view is based on the conclusions of the private language argument and the sections after it about mental phenomena. Wittgenstein now applies these concerns to verbs like "to believe, "to know," "to be convinced," "to expect," "to recognize," and so on.

None of these can accurately be said to describe a mental "process." When I sit down in a chair, I believe that it will not break. Saying this is perfectly in order. But does the idea of the chair breaking as a possibility ever actually occur in my mind? No, of course not. If someone asked me whether I recognized my desk when I entered my room in the morning, the answer would be yes. But I certainly did not have any experience of "recognition" upon seeing it.

Nor can we say that these verbs describe "feelings." William James, for example, proposed that we have a number of subtle, phenomenologial tones and hues of experience that we have an impoverished language for. States like "believing," "being convinced," and so on may have associated feelings that are ineffable and difficult to capture in words but are only known internally. But this is just the private language argument rearing its head again.

I might say "I FEEL convinced that...," but that does not mean that there is an ineffable feeling tone that wells up within me when I am "convinced" of something. Wittgenstein asks if it would be possible to have a deep feeling of ardent love or firm hope for only one second, no matter what preceded or followed it. This proves very difficult to imagine. And that is because these feelings do not describe any amorphous inner state. They describe patterns of life that emerge in our language-games. They are explained not by what ACCOMPANIES them in the moment but what SURROUNDS them in time and space.

When we try to capture something like "understanding" as a feeling or an atmosphere, we are looking for the wrong kind of explanation. It is similar to trying to describe the aroma of coffee. William James would say that we lack a language for but have an ability to identify and comprehend the specific aroma of coffee in a phenomenological way. Wittgenstein thinks that this whole idea is just a mistaken application of rules of different language-games. [25]

We can "describe" the aroma of coffee in many ways (these examples and framing are diectly taken from P.M.S. Hacker):
"It smells fresh" (or rich, faint, delicious...)
"It is THIS smell" (spoken in a coffee shop)
"It is the smell when beans are roasted"
"It smells similar to chicory"
These are all perfectly adequate "descriptions" in our language-game. And yet, we still feel that the "real essence" of that smell is left out. None of these can explain it to someone who completely lacked the sense of smell altogether, for example. For Wittgenstein, this "perfect" description could never exist. And it is NOT because the phenomenal experience of a smell is mysterious, ineffable, and too fine for our crude language to capture. It is because we are simply confused about the rules of our language-game if we try to "describe" it.

4.o. §611-628: Free Will

We have a distinction between agency and passivity in our language-game. This goes beyond the mere distinction between animate and inanimate objects. We know that we conceptualize the action of a person who writes down someting on a blackboard differently from the action of an acid that eats away at a metal. But we also talk differently of an acid when it is dissolving a metal versus when it is being made to change by an alkaline solution being added to it. We distinguish what a ship does when it is careening over the waves full speed ahead versus when it is being made to drift along the waves helplessly. [26]

Plants and animals can certainly be said to be passive or active as well, but here we can speak more sensically of them having "goals" or "aims" to their action. A plant grows in order to have access to sunlight and water. A dog can jump on a table in order to get at some piece of food. When a goal or action exists, we traditionally apply the idea of a "will" to the performer of that action. Wittgenstein engages with the question of the "will" here with an allusion to Schopenhauer:

Witgenstein asks what is is that our "willing" apparently consists of when we act of our own "free will." What distinguishes humans is that we can reflect on our reasons for acting, entertain "reasons" for acting, and entertain alternate courses of action. We can "justify" our actions. Of course, this all presupposes a language-game. We tend to distinguish actions that we do voluntarily from ones that we do involuntarily. We say that WE are the ones who write or eat or talk, but not that WE are the ones who make our hearts beat or our pupils dilate in darker environments. We say that WE move our leg when we kick it during a martial arts lesson, but not when the doctor tests our reflexes by hitting our knee with a small hammer.

These are clearly very different language-games. But where do we actually locate that difference? Do we look "inward" and find some experience called "willing" before we choose to move our leg? The fact is that the vast majority of our body's actions happen without us thinking, even in the subset that we would consider "voluntary." We do not "think about" scracting our head before it itches. We simply scratch it.

We appeal to something called the "will," but what could this be? Is it a kind of wish or desire we have? Certainly "wanting" to do something does not immediately transfer to action, so this does not make sense. If the will is some immaterial substance inside of us, it seems hard to explain how it causes the physical matter of our body to behave in one way or another. It turns into a kind of magical psychokineses. Is the "will" merely a pattern of action in our brain's neurons? If so, then nothing in the end distinguishes it from our heart beating or pupils dilating, which is exactly the kind of thing we wanted to define it in contrast to in the first place!

Wittgenstein's next move is predictable: He wants to attack this idea of the "will" or "intention" by showing us how our language works. But this section seems incomplete. What he ultimately seeks to show is that we cannot call "willing" any kind of experience. "Willing" is not an action and for that reason cannot be said to be either voluntary or involuntary. When we raise our arm, we raise our arm. We do not "bring about" the raising of our arm. The will is an attempt to put some intermediary in the action where none is needed to explain it.

Wittgenstein ultimately seems to conclude that the difference we have between sentences like "I raised my arm" and "My arm raised" occurs as a distinction between actions that are UNSURPRISING and ones that are SURPRISING. Can one imagine saying "My hand scratched my nose because it itched?" in an ordinary conversation? As usual, it becomes the surrounding context that explains everything we need in the use of these words.

4.p. §629-660: Intention

An action with "intention," put most plainly, is any action for which we can ask the agent his reasons for acting. We don't ask someone why he sneezes or yawns. Nor do we ask him why he taps his fingers when waiting. Like other verbs such as "to think" and "to believe," saying that X "intends" to Y can cause us to adopt a misleading philosophical picture. This treatment also appears incomplete. It focuses on one example in particular: When we are speaking, get interrupted, and then resume speaking and say something like "I was going to say..."

More broadly, we can think of this at the moment where we have been interrupted or stopped ourselves just before we do sometihng, i.e. carry out what was our INTENTION. A treatment of this specific case ends up being particularly illustrative about the problem of intention generally.

When I am interrupted and then remember what I was going to say, what am I remembering? Our intention, we think. But what does this consist of? We can remember the context just before we were about to speak and the feelings, impressions, etc. surrounding it. But the sum of these does not seem to produce something called an "intention." We refer to a point in time when we remember that moment, perhaps, but there is no "experience" to find here. All the potential "experience" vanishes the further we look. It's not like remembering that we left the oven on, where we could have distinct images in our mind of not turning the button off, leaving the kitchen in a hurry, locking the door, exiting the gate, and so on.

We can only understand the real answer as to what goes on here if we stop looking for an "explanation" at all. We want to ask how it is possible for us to remember what we were going to say when there is no thing in our head called an "intention" nor any real "content" to the event or experience of the moment in which we were about to say it. We cannot explain this in the same manner by "grounding" it in some other piece of phenomena. But it does not need one either.

All we are doing when we make the statement "I was going to say..." are re-orienting ourselves and directing ourselves back towards something. We not need to appeal to anything that goes on in our heads here, any more than we do when we hear a noise, look in the direction of it, see nothing, and then re-orient ourselves back to what we were engaged with.

The language-game of "I intend..." is meant to herald and call attention to actions that follow. To use Hacker's example, when I'm teaching a child how to play a game, I will say "Okay, I'm going to throw the ball now!" before I throw it. When we say "I intend to...", our use is essentially the same, just in a much more culturally complex language-game. In simple terms, it is essentially an elaborate method of signaling our actions. [26]

4.q. §661-693: The Meaning of "Meaning"

The last broad theme we see in the main section of the Philosophical Investigations is about the verb "to mean" and what happens when we "mean" something or else by our words. At first glance, it appears as if "meaning" is a mental activity that accompanies us when we "mean" something. We typically use it in contrast to our speech versus, say, that of a parrot, which repeats words without this psychical component. But here we can rehearse some of the argument for "intending" to cast doubt on the idea that there is some "experience" or "content" to the "act" of meaning, insofar as it can be called an act at all (which it really cannot).

"To mean" makes no sense as an action verb. You cannot start to "mean" something at a specific time or be interrupted in the middle of "meaning" something and start to "mean" it again. You cannot forget how to "mean" or find it difficult to "mean" or enjoy "meaning." You cannot "mean" quickly or slowly or with any other adverb. We might think that meaning consists in having a kind of mental image of the "meant" thing in our mind when we speak, but this clearly does not have to be the case. We can tell someone to shut the door while our only thoughts are about how cold it is getting in the room. And yet it would be absurd to say that I did not "mean" for him to shut the door. [27]

Wittgenstein gives the example of being in pain while a piano is being tuned loudly in the next room. I may be in pain but the dissonant noises are distracting enough that I'm more focused on them than my pain. If I said "It'll stop soon," would we think that I must really mean the piano noises and not my pain just because I was more focused on the piano noises? The exact purpose of this thought-experiment is a bit hard to follow, but I interpret it as showing that "meaning" cannot simply be a matter of focus or attention any more than it can be a matter of mental imagery.

Most of the intuitions we have about meaning are correct. We cannot "mean" a nonsense. We do not "mean" things that we repeat automatically like a parrot. We say something and we "mean" it if we want the other person to understand what we say. But all this really amounts to is that to know the "meaning" of a word, one must know how to use it correctly, respond to its use, explain its use, and so on. One must master its use. That is all that there can be to "meaning" something: To adequately use language within one's form of life.

4.r. Part II. Seeing an Aspect

The Philosophical Investigations ends with a section appended on the end called "Part II." But this name makes it sound much longer than it actually is. It is an even more fragmentary and spotty collection of notes. Many of the themes are repeated from the main text. The one notable addition is what is called "seeing an aspect." What is the difference in the verb "to see" when we say "I see a hot air balloon," such as when we take an eye exam, and when we say "I see a likeness between these two faces?"

Clearly, something different is happening here. We can fail to "see" a likeness btween two faces. But clearly if this happens it is not because of a physical defect in our vision. There's nothing involved in this "seeing a likeness" that can be translated into a sensory experience. Someone who fails to see the likeness and someone who succeeds at it do indeed "see" in precisely the same way from a physiological perspective. We can also certainly find no objective difference in the object iteslf, as this discrepancy could emerge even if two people saw it under the exact same circumstances (i.e. the lighting and angle of the picture were the same for both viewers).

We are tempted, then, to explain that the difference must occur in our brain. We tend to think of our visual impression of the thing "detaching" from it like a membrane and being absorbed and morphed in our brains. But this does nothing to explain the discrepancy. Earlier sections of this book should make us skeptical of the coherence of the idea of private mental images at all. But even if they did exist, it does not solve the problem. We still cannot explain exactly what it is that changes in how we perceive this object while still making it connected enough to the original picture to not be a mere hallucination.

The answer comes once we pay attention to the special grammar of the term "seeing as," which is something very distinct from pure visual perception. Wittgenstein looks at a famous optical illusion, the duck-rabbit, which can be seen as either a duck or a rabbit:

Let's say that a hypothetical subject only ever saw this picture in the context of rabbits. He saw it in books about rabbits, next to other pictures of rabbits, and so on. If we asked him "What do you see?," he certainly would not say "I see it AS a rabbit." He would just say "I see a rabbit." But if someone else did in fact understand the ambiguity of the picture and responded with "I see it AS a rabbit, the difference could not be found in terms of what they are seeing visually. The concept of "seeing as" should not be understood as something that is caused IN us by something "out there" in the world. It has to do with how we REPSOND to what we see:

The latter example here is clearly distinct from the first one. But we can only explain the difference between these two in the context of behaviors surrounding them. The truth is that "visual experience" is not something that can be detached from all other forms of life. Our action of seeing does not come before our action of responding to what is seen. It is intimately bound up with our verbal descriptions, our gestures, our comparisons, and so on. [28]

Why does a smiling face turned upside-down look more uncanny than an abstract figure turned upside down? Because it is harder to mimic it, to describe it, or to imagine talking it. Why does a word reversed as if in a mirror look more uncanny than an abstract shape being reversed? Because it is harder to read, to copy, or to imagine seeing on a page. Visual experience itself is bound up with a nexus of behaviors and responses that we learn and grow within, that is to say, our form of life.

The remainder of this section has a number of remarks that elaborate on this theme and others. Many are quite thought-provoking and many still quite difficult for me to interpret. I present some of my favorites here as an end to this lecture:

FOOTNOTES

1. Ludwig Wittgenstein [trans. David Pears & Brian McGuiness], Tractatus Logico Philosophicus 6.53, Routledge Classics, 2001

2. Marie McGinn, The Routledge Guidebook to Wittgenstein's Philosophical Investigations, Routledge, 2013, p. 33-69

3. Ibid.

4. Gordon P. Baker & P.M.S. Hacker, An Analytical Commentary on the Philosophical Investigations, Volume 1: Wittgenstein: Understanding and Meaning, 2009, Wiley-Blackwell, p. 107-201

5. Ibid.

6. Ibid., p. 201-244

7. Marie McGinn, The Routledge Guidebook to Wittgenstein's Philosophical Investigations, Routledge, 2013, p. 73-106

8. Ibid., p. 113-134

9. Gordon P. Baker & P.M.S. Hacker, An Analytical Commentary on the Philosophical Investigations, Volume 1: Wittgenstein: Meaning and Mind, 1993, Wiley-Blackwell, p. 3-147

10. Marie McGinn, The Routledge Guidebook to Wittgenstein's Philosophical Investigations, Routledge, 2013, p. 113-134

11. Gordon P. Baker & P.M.S. Hacker, An Analytical Commentary on the Philosophical Investigations, Volume 3: Wittgenstein: Meaning and Mind, 1993, Wiley-Blackwell, p. 3-147

12. Marie McGinn, The Routledge Guidebook to Wittgenstein's Philosophical Investigations, Routledge, 2013, p. 143-170

13. Gordon P. Baker & P.M.S. Hacker, An Analytical Commentary on the Philosophical Investigations, Volume 3: Wittgenstein: Meaning and Mind, 1993, Wiley-Blackwell, p. 171-272

14. Ibid., p. 287-335

15. Ibid.

16. Ibid.

17. Ibid., p. 392-423

18. Marie McGinn, The Routledge Guidebook to Wittgenstein's Philosophical Investigations, Routledge, 2013, p. 177-189

19. Gordon P. Baker & P.M.S. Hacker, An Analytical Commentary on the Philosophical Investigations, Volume 3: Wittgenstein: Meaning and Mind, 1993, Wiley-Blackwell, p. 468-494

20. Marie McGinn, The Routledge Guidebook to Wittgenstein's Philosophical Investigations, Routledge, 2013, p. 177-189

21. Gordon P. Baker & P.M.S. Hacker, An Analytical Commentary on the Philosophical Investigations, Volume 3: Wittgenstein: Meaning and Mind, 1993, Wiley-Blackwell, p. 513-571

22. Gordon P. Baker & P.M.S. Hacker, An Analytical Commentary on the Philosophical Investigations, Volume 4: Wittgenstein: Mind and Will, 2000, Wiley-Blackwell, p. 1-49

23. Ibid., p. 49-69

24. Ibid., p. 69-101

25. Ibid., p. 111-157

25. Ibid., p. 191-239

26. Ibid., p. 239-261

27. Ibid., p. 261-285

28. Marie McGinn, The Routledge Guidebook to Wittgenstein's Philosophical Investigations, Routledge, 2013, p. 189-204