Wittgenstein: When a Philosophical Question Arises from Language

Imagine a hypothetical conversation. One person says, “You don’t understand me.” The other replies, “I understood every word perfectly.” The first wants their experience to be taken seriously. The second reports being able to retell the content. They use one word but perform different actions. After several repetitions, the question becomes almost philosophical: what does understanding another person mean at all?

We could answer with a large theory of consciousness. But we could first find out what is happening with “understand” here. Ludwig Wittgenstein proposes this turn of attention: rather than rushing to find a hidden object behind a familiar word, examine its uses. Sometimes the difficulty remains. Sometimes we discover that the dispute was organised by a question that quietly joined several different tasks.

What this article explores

This article concerns the later Wittgenstein and his way of noticing philosophical difficulties in language at work. We will examine meaning as use, language-games, family resemblance, and the problem of following a rule. We will then discuss why the private-language argument does not abolish personal experience and why returning to ordinary speech does not prohibit science or complex thought.

Our contemporary scenes are hypothetical examples, not episodes from the philosopher’s life. They show the difference between clarifying words and resolving the situation itself. Clarifying “respect” does not yet secure respectful treatment. Understanding “freedom” does not yet change restrictive conditions. A more precise question can, however, reveal what participants actually disagree about. We will preserve that boundary and resist calling every difficulty a linguistic misunderstanding.

1. Why a familiar word can become a trap

A familiar expression does not guarantee a clear question. In everyday life, words enter particular actions. In philosophical reasoning, they are sometimes extracted from those actions and required to reveal the single thing they supposedly always designate.

Near the beginning of Philosophical Investigations, §§1–2, 11, 23, Wittgenstein examines the limits of picturing language as a collection of names for things. His examples show words doing varied work: naming, requesting, reporting, supposing, and thanking are different forms of participation in activity. Naming matters, but does not exhaust language.

Take our own example: “I promise.” We can ask which object “promise” names. Yet a promise matters as more than a report of an inner state. It changes expectations between participants. “I promise” on a theatre stage, in a signed agreement, and in a conversation with a friend does not create identical obligations.

This does not prove that the word has no meaning. Rather, its meaning becomes clearer through differences in the situation: who speaks, to whom, in what action, and under which rules the statement is accepted. Removing the word from these relationships tempts us to invent an invisible object, “the promise itself,” which must then explain every case.

Our reasoning does not propose abandoning general concepts. It proposes checking what we expect from them. “What is a promise?” may ask for an explanation of a word, a legal criterion, an assessment of moral duty, or an investigation of the experience of obligation. Until the tasks are separated, any answer may seem both correct and inadequate. Sometimes the obscurity lies in the question rather than in its subject’s mysterious depths.

2. Meaning as use, with an essential qualification

In §43 of Philosophical Investigations, Wittgenstein connects a word’s meaning with its use for a large class of cases, explicitly retaining a qualification: this does not apply to all uses of “meaning.” The famous formulation can easily become an absolute slogan that loses the text’s own caution.

Consider the hypothetical word “done.” A report’s author says it after finishing a draft. An editor says it after checking. A client says it after accepting the result. Each may use the expression normally. Trouble arises if participants assume the same finishing criterion without agreeing on it.

How can we check the meaning? We can ask which action becomes permissible after “done.” Sending it to colleagues? Paying? Starting to use it? Publishing it? This reveals consequences of the usage and helps make its criterion explicit. This is our contemporary example; the philosopher did not discuss these editorial stages.

Meaning must be distinguished here from personal desire. One participant may want “done” to mean full acceptance, but that does not establish a shared rule. Nor is meaning reducible to statistics of frequent usage. If many employees habitually say “checked” after a cursory glance, the custom alone does not make their conclusion reliable.

Attention to use is therefore compatible with criticising a practice. We can establish how an expression currently works and then decide whether that work suits our purposes. Clarifying an existing rule and proposing a new one are different actions. Replacing the first with the second makes it easy to speak as though the desired agreement already exists. Attentive reading restores the difference: understanding a use does not automatically mean approving it.

3. A language-game need not be unserious

Even pointing at an object does not eliminate context. In §§28–30, Wittgenstein considers how an ostensive explanation depends on a word’s role. Primary text. Continue with our example: a teacher points to a button on a panel and says “start.” A pupil may understand the part’s name, the action it initiates, or permission to press it right now. If prior learning has not established the task, the gesture remains ambiguous. Adding a label helps only when someone knows how to use labels in those circumstances.

This is not an argument against definitions. An explanation can be entirely sufficient for one listener and unsuitable for another. Difficulty begins when sufficiency is taken to mean independence from every circumstance of application. In our case, a new pupil needs the sequence of actions demonstrated; an experienced colleague needs only one word. The difference does not necessarily reveal deeper inner understanding in the second. They already know the practice in which a short explanation works.

Checking understanding may therefore require another case rather than another definition. Can the pupil distinguish starting a test from starting the device? Do they know when the button must not be pressed? Such checking does not exhaust their thinking, but reveals a significant part of their command of the expression. Conversely, correctly repeating a formula does not establish readiness to act. The difference between a dictionary answer and application becomes visible without guessing what is hidden inside another person’s head.

Wittgenstein’s “language-game” emphasises the connection between speech and activity; §23 lists very different actions. “Game” does not declare everything entertainment or suggest that statements have imaginary consequences.

In our hypothetical conversation, “You can leave” may be permission, a threat, advice, or a statement of physical possibility. A dictionary helps translate individual words, but does not yet establish the action occurring. Tone, the preceding conversation, the participants’ relationship, the possibility of refusal, and subsequent behaviour matter.

This does not entitle us to guess someone’s intention from a convenient version. On the contrary, varied uses increase the need for information. When context is unknown, an honest answer sometimes offers several possible readings. “Here, it could have been permission” is more precise than “he definitely wanted to set you free.”

It also helps to distinguish explaining a language-game from participating in it. An observer may describe negotiation rules without supporting a deal’s terms. Someone can understand a threat and reject it. Understanding an action does not equal agreeing to it.

Words can move between activities. We speak of “debt” in accounting and gratitude, of “error” in calculations and relationships. Sometimes the transfer is fruitful. Sometimes it creates a false demand: as though gratitude must be repaid in an exact equivalent, or a human disagreement corrected by one right answer.

A language-game is useful as a question about an action’s structure: what counts as a move, reply, mistake, or ending? This question does not diminish a situation’s seriousness. It helps us see the expectations speech creates. In conflict, it may clarify the difference between a request, a demand, and a judgement even if the parties continue disagreeing afterward.

4. Family resemblance and the search for one essence

In §§65–67 of Philosophical Investigations, Wittgenstein invites us to look at intersecting similarities between games rather than demand in advance one feature common to all. “Family resemblance” names this network of overlaps. The move does not prohibit precise definitions where they are needed.

Take our example of “support.” Support can come through money, attention, technical help, a public statement, or respect for a refusal. These actions resemble each other in some respects but need not involve the same sequence. Not giving advice does not prove an absence of support. Giving abundant advice does not prove its presence.

It does not follow that anything can be called “support.” Examples, distinctions, purposes, and the possibility of discussing boundaries remain. In a hypothetical situation, persistent interference may be presented as support even though its recipient asks for it to stop. Appealing to the concept’s imprecision does not erase that particular refusal.

Family resemblance helps us avoid confusing the absence of one shared definition with the absence of distinctions altogether. We may recognise many cases while lacking a short formula that decides every new one in advance. Precision sometimes comes through comparison: this action resembles support in intention but differs in respect for boundaries and in consequences.

A specially established criterion may be needed at work. An employee support programme, for example, should clearly define available services. An organisation may draw a firmer boundary for that particular task. It should simply avoid presenting the definition as the discovered eternal essence of all human support. It resolves a practical question in a chosen context; it does not abolish the word’s other uses.

5. Following a rule is more than holding a formula in mind

A rule seems reliable until we ask what makes its application correct. A written formulation may not suffice: people need to know how to continue, correct mistakes, and distinguish compliance from its appearance. Competing interpretations surround Wittgenstein’s discussion; the Stanford Encyclopedia, §3.5, warns against presenting the sceptical reading as the only one without qualification.

Our hypothetical example is an instruction to “process requests in order.” One employee understands the order as time of arrival. Another first separates urgent requests. A third groups requests from one customer. All may sincerely believe they are continuing an established practice. The formulation does not explain which practice was intended.

To advance the discussion, we need ordinary and borderline examples. What if two requests arrive together? What counts as urgency? Can a request with missing information be skipped? Rules become intelligible through application and correction rather than endlessly adding explanations to explanations.

There is a limit to detail. Trying to anticipate every possibility makes an instruction unreadable while still requiring fresh judgement. That is no reason to abandon rules. It is a reason to establish what training and what procedure for disputed cases enable people to work with them.

Who may declare a mistake also matters. An employee may act consistently yet face a manager who changes the criterion retrospectively. The problem then concerns power and fairness within the practice, as well as textual interpretation. Our example extends the philosophical question without attributing the extension to Wittgenstein. Clarifying rule-following can sometimes reveal that an organisation lacks a stable way to distinguish a rule from an arbitrary demand.

6. Private language and the right to personal experience

Discussion of private language can be read too quickly, as though someone cannot know their own pain or only what others see counts as true. Neither conclusion follows from the question under examination. The academic overview, §3.5, concerns distinguishing correct from incorrect application of a sign meant to designate a sensation accessible exclusively to one speaker.

In our hypothetical example, someone keeps a personal diary that nobody else reads. This is ordinary private use of shared language: the writer knows how words are used, can compare entries, and can detect inaccuracies. A diary’s secrecy does not itself create a philosophical private language.

Another thought experiment works differently. Someone introduces a sign for a sensation and treats their current impression of applying it correctly as the sole criterion of correctness. A question arises: what would distinguish error from correct recognition here? Our example illustrates the problem of a criterion; it does not replace the philosopher’s entire argument.

For human communication, it is essential not to turn this problem into distrust of self-report. “I am in pain” has its place in life. A listener need not acquire the same sensation to take the statement seriously. Discussing meanings does not authorise them to declare another person’s experience invented.

Taking something seriously and understanding it completely are nevertheless different. Someone can acknowledge another’s experience, not know its intensity, and be wrong about its cause. Attentive speech preserves all three points. We might say, “I hear that this is difficult for you; I don’t yet understand exactly what happened.” This joins respect and uncertainty without promising direct access to another’s inner world.

7. What philosophical clarification means

Publicly accessible criteria do not mean the majority is always right. Consider our thought experiment: participants use a shared measurement scale and obtain the same number, but the instrument turns out to be faulty. The practice allows us to understand what they asserted; further checking reveals the error. Meaning and truth differ. Agreed ways of speaking make disagreement possible; they do not decide its winner beforehand.

We can thus state the concern about attention to ordinary language more precisely. If ordinary usage is treated as the final authority on truth, it shields habit from checking. Yet describing what someone means does not endorse what they say. We can fully understand an accusation and reject its grounds. We can understand a promise and doubt its feasibility. Clarity allows a sharper question; it does not replace the answer.

Sometimes new knowledge changes an existing vocabulary. Suppose researchers discover a difference between two phenomena previously called by one word. They will need a new term or a refinement of the old one. Requiring a return to former usage would obstruct inquiry. But the introduced term also needs explaining: which cases it distinguishes, which error it makes noticeable, and where it does not yet apply. Specialised language becomes useful through this work, rather than through an unusual name alone.

In §133 of Philosophical Investigations, Wittgenstein speaks of multiple methods and compares them to different therapies. This describes work on philosophical difficulties; it does not turn the text into medical or psychological care. We use “clarification” here to preserve the difference between those tasks.

Take the hypothetical question “Am I really free if every decision of mine has a cause?” Sometimes it calls for serious metaphysical discussion. Sometimes it mixes different meanings of freedom: absence of coercion, available alternatives, independence from any cause, or the ability to act on one’s decision. Before arguing for an answer, it helps to notice the mixture.

If someone says “I am not free” because they cannot refuse an imposed schedule, investigating an uncaused will may not answer their question. If causation’s relationship to action is what interests them, describing a work schedule cannot replace a philosophical argument. The right scale is determined by the task, rather than by the prestige of greater abstraction.

Clarification may change a difficulty’s form. “Does freedom exist at all?” becomes “what conditions would make this decision mine?” “Is complete understanding possible?” becomes “what do I want the other person to acknowledge or be able to retell?” These are our examples of a question’s movement, not a compulsory Wittgensteinian procedure.

Sometimes the new question is easier, sometimes harder. It may reveal a real divergence of purposes previously hidden behind a shared word. Participants may understand “justice” identically yet defend incompatible ways of distributing a resource. Successful linguistic clarification then means a more precise dispute, rather than its disappearance.

8. An objection: language does not resolve everything

A strong objection to overextending this approach is obvious: real difficulties can be material, political, and bodily. Someone lacks resources, has been denied access, or faces danger. Establishing the use of “problem” does not remove those circumstances.

Imagine a hypothetical residents’ meeting. They argue about who must pay for repairs. One proposes starting with responsibility as a concept; another presents a damaged pipe and an estimate. Both actions may be needed. But endless discussion of the word becomes avoidance if the evidence and obligations are already sufficiently clear.

Clarification is useful to the extent that it reveals a significant distinction. If a criterion, the subject of disagreement, or an available choice changes after clarification, work has been accomplished. If participants merely exchange one abstract expression for another, its value remains doubtful.

Ordinary usage can also become pressure. “Nobody talks that way” may serve to ignore new experience or justify an unsuccessful norm. Language allows new concepts, changes of rules, and descriptions of things for which no convenient word previously existed. Examining current usage does not prohibit change.

Our practical reading therefore includes a critical question: is clarification protecting the present order? Someone asking for humiliating treatment to stop need not first supply a perfect definition of humiliation. Evidence of error cannot be dismissed solely because it does not fit a familiar vocabulary. Attention to language should support precision rather than become another obstacle to reporting what happens.

9. What this means in practice

Finally, we should distinguish a philosophical question from one that only resembles it in form. “What is memory?” may request an account of biological mechanisms, ask whether a recollection is reliable, or express a difficulty: why do I call a present representation a memory of the past? The first needs research, the second information about a particular event, and the third clarification of the concept’s workings. One general answer risks satisfying none of the tasks.

We need not choose only one form of work in advance. Sometimes the tasks connect: a scientific discovery forces reconsideration of a familiar distinction, while conceptual refinement helps formulate a testable question. Yet connecting them requires first distinguishing them. Otherwise, experimental results are offered as answers to a word’s meaning, and a definition as proof of a mechanism. Reading Wittgenstein attentively invites us to check our own question’s movement: at what step did I stop asking about what I began with?

We can start with one word around which a dispute recurs, rather than all our beliefs at once or an attempt to tell someone they are “confused by language.” That accusation already creates a new problem: one participant claims the right to determine meaning for the other.

Instead, in a hypothetical situation, we can describe different actions. “When I say ‘understand,’ I want us to agree about what happened.” Or, “For me, understanding right now means you can retell why this matters.” These formulations need not win the dispute. They merely make one’s own request distinguishable.

We can then check an example and a counterexample. What would respect look like in this conversation? Which action would remain respectful even if the other person disagreed? And which agreement would be disrespectful if secured through pressure? Such questions help reveal whether we have identified respect with an outcome convenient to ourselves.

It helps to distinguish differences in words from differences in interests. If both parties understand “by Friday” alike but one needs the result sooner, the problem is not the date’s meaning. If they agree about what happened but judge its permissibility differently, clarifying the facts again may add nothing.

Finally, some expressions may appropriately remain open. Someone is entitled to say they cannot yet name an experience precisely. This neither makes their speech meaningless nor obliges them to accept the first offered label. Clarification may include examples, a pause, a partial formulation, or an honest unknown. Its purpose is to make thought more accessible without demanding completion at any cost.

10. Life remains after clarification

We should also distinguish not understanding a word from struggling to imagine its application. Someone may know “forgive” but not see what it would mean after a particular event: restoring contact, renouncing revenge, changing a judgement, or something else. Demanding a dictionary definition will not finish that work. In our reading, comparing possible actions and naming which ones the speaker considers is more useful. The philosophical example must not quietly become a demand to forgive. Clarifying the concept leaves participants free to reject a proposed action. It makes a choice distinguishable rather than choosing for them. This is another case in which clarity of use and desirability of an outcome need different grounds.

Return to the hypothetical conversation about understanding. One participant can now say, “You retold the event correctly, but I wanted you to acknowledge its importance to me.” The other can reply, “I acknowledge its importance, though I don’t yet share your explanation of its cause.” Agreement becomes more precise, and disagreement gains a particular subject.

This ending does not resemble a philosophical triumph. The conversation may remain difficult. Participants may need time or a new decision. But “you don’t understand” no longer has to encompass the entire difference between them.

A Wittgensteinian reading teaches attention to the work a question performs. Sometimes it asks for a fact. Sometimes a definition. Sometimes agreement, justification, acknowledgement, or permission. If these actions merge, one answer must perform an impossible task. Even a careful theory then seems indifferent because it answers another request.

For us, a modest possibility follows: before making a large conclusion, notice the key word’s use and examine particular cases. This is no universal way to resolve every conflict. Sometimes it shows that further philosophising adds nothing and action should be discussed. Sometimes, conversely, it helps find a serious philosophical question deserving a full argument.

Words can impede understanding, yet words also communicate distinctions and correct mistakes. Attention to language does not lead away from the world into a dictionary. At its best, it returns us to what happens: who said what, which response they expected, what became clear, and what remains unresolved. After clarification, life remains, described a little more precisely, with the costs of its decisions and the participants’ right to disagree.

Sources used

1. Ludwig Wittgenstein. Philosophische Untersuchungen, §§1–2, 11, 23, 28–30, 43, 65–67, 133. German primary text: varied uses, ostensive explanation, the qualified formulation of §43, family resemblance, and multiple methods. The link leads to an electronic publication; full reading of the work is not claimed.
2. Anat Biletzki, Anat Matar. Ludwig Wittgenstein, Stanford Encyclopedia of Philosophy, §§3.3–3.7. Academic overview used for a cautious account of rule-following, interpretive disagreements, and private language.