Table of contents
Prelude
There is a peculiar kind of argument that does not defeat a question. It diagnoses the desire to ask it, describes the scenery around it, then walks away as though the investigation has been completed.
A wonderfully efficient arrangement. The questioner supplies the question, the philosopher supplies the illness, and nobody has to supply the proof.
I kept running into that move whenever somebody asked whether the things called games have anything in common. Family resemblance would arrive before the candidate did. Eventually the phrase itself stopped bothering me, but the confidence surrounding it did not.
I was not yet sure whether I was annoyed with Wittgenstein or with the little Wittgensteins who appear whenever somebody tries to define anything. So I went back to the actual neighborhood: the interlocutor in Philosophical Investigations, the games in sections 65 through 71, the remarks on exactness that follow them, the conception of philosophy around sections 89 through 133, and the earlier diagnosis of a “craving for generality” in The Blue Book.
I have no philosophical team to defend here. If anything, I lean toward the thought that human beings do not arrive with an essence waiting to be obeyed. That tells me nothing in advance about what games, numbers, laws, or any other objects have in common. A temperament is not a proof either.
That made the position harder to dismiss, and if an objection survives only while its target remains unread, it deserves to die before the target does.
Then I read what he was actually doing
The question in section 65 is not dropped from the sky. Wittgenstein has been describing many language-games, and an interlocutor objects that he has still not said what is common to them, what makes them language or parts of language. The objection asks for the general form. It asks for essence.
The answer begins before section 66. Wittgenstein says these activities have no one thing in common that makes us use the same word, but are related in many ways, and those relations explain the shared name. Then comes the famous instruction not to think but to look. Board games, card games, ball games, winning, losing, skill, luck, amusement, competition: every proposed likeness gives way somewhere while another appears elsewhere. Section 67 names the resulting network “family resemblances”, and the strength of its thread comes from overlapping fibers rather than one fiber running through the whole length.
He even anticipates one obvious reply. Someone may say that the constructions share a disjunction assembled from their overlapping properties, or that continuous overlapping itself runs through the thread. Wittgenstein calls this “playing with words”. Perhaps it is, if the proposed commonality merely restates an extension and explains nothing. But that verdict already depends on what common is being asked to do. Being useless as an essence does not make it nonexistent as a relation, and calling the distinction wordplay does not pay for erasing it.
By section 69, explaining the ordinary word “game” no longer requires an analytic boundary. We show examples and indicate that similar things belong with them. Sections 70 and 71 resist the complaint that a concept must be sharply bounded before it can be a concept at all. The discussion then keeps going: a sample becomes general through its use, knowing what a game is can be expressed through examples and analogies, and a sharp boundary drawn later produces a related concept rather than uncovering the blurred one that ordinary use had secretly contained.
Read in place, this is not a lazy attempt at taxonomy. Wittgenstein is attacking a picture of meaning. The picture says that a general word must receive its unity from one common property, that understanding the word consists in possessing that property, and that an unclear boundary reveals an incomplete concept. His reply is that our use of a word can be ordered, teachable, and perfectly serviceable without an analytic definition waiting behind it.
That point matters. A blurred boundary is not the same as no use. A rule can be adequate for one purpose without settling every possible edge case. We can know how to continue in familiar circumstances while remaining unable to state a formula that mechanically decides every future application. Ordinary language does not wait for philosophy to finish its paperwork before anybody can speak.
On that much, I have very little desire to argue with him.
Heather Gert argues that family resemblance need not constitute membership at all. It may instead provide fallible evidence for membership in an extension grounded elsewhere. That reading protects Wittgenstein from having to turn resemblance into a membership algorithm, but it narrows what the metaphor can establish. Evidence may guide recognition without proving that no other account or commonality exists beyond it. A constitutive reading would owe the resemblance relation, while Gert’s evidential reading owes no such thing and receives no ontological authority in return. Either way, an inspection of selected similarities has not earned the right to decide what inquiry may continue.
The stronger it gets, the more it owes
Family resemblance is only one piece of a broader philosophical method. The Blue Book made that impossible to ignore.
There, the “craving for generality” gathers several tendencies together. One is the search for something common to everything covered by a general term. Another is the temptation to imagine a general mental picture behind our understanding of such a term. A third confuses a hypothetical mental mechanism with a state of consciousness. The most important one for my irritation is the attraction of scientific method: reducing phenomena to a small number of laws and unifying different cases through generalization. Wittgenstein says that philosophy is tempted to imitate this method and walks into darkness. Philosophy, in his formulation there, is “purely descriptive”.
Later in the same book, however, he calls philosophy, “as we use the word”, a fight against the fascination of forms of expression. He even says the old title may be both appropriate and misleading, as though his subject were an heir to philosophy rather than its uncontested owner. That wording sounds local, almost stipulative. It describes the work he has chosen and warns that the name itself may smuggle in the wrong inheritance.
The issue returns in Philosophical Investigations through a more specific target: the sublime purity once demanded from logic, including the picture Wittgenstein had drawn in the Tractatus. Section 104 warns that we predicate of a thing what lies in our method of representing it, then mistake the convenience of comparison for a state of affairs at the highest generality. It is difficult to ask for a cleaner warning against method turning its own field of vision into the shape of reality.
Then the text says philosophy is not supposed to advance hypotheses. Explanation gives way to description in section 109, and philosophy leaves the actual use of language as it is in section 124. The relevant facts are already open to view, but their familiarity makes them hard to notice. Section 121 lets philosophy speak about the word “philosophy” without inventing a second-order discipline, much as orthography can deal with its own name. Sections 130 through 132 present language-games as objects of comparison, models as measuring rods rather than preconceived laws, and each order as one possible order serving a particular end. By section 133, there is no single philosophical method, but different therapies demonstrated through examples. The desired discovery gives philosophy peace.
So the method does examine itself. It describes its procedure, its purpose, the danger of its models, and even the temptation to confuse representation with reality. What it does not make equally clear is the edge of its competence. A method can face a mirror without learning which wavelengths never reached the glass.
Philosophy has become a clinic. That is fine, provided somebody remembers that diagnosing the patient is still a claim about the patient.
The strongest Wittgensteinian reply was now obvious. Demanding a scientific proof from this project misses the project. Family resemblance is not a rival scientific theory about the hidden ontology of games. It is part of a grammatical investigation into how words work and how philosophy becomes trapped by a false demand placed upon them. “Look and see” is not an induction pretending to be a theorem, but therapy for the urge to produce one.
That reply is fair, which is inconvenient in exactly the useful way, but it also draws the boundary I need.
Rejecting scientific method does not reject the obligation to prove what the method claims. The proof need not be an experiment, an induction, or a formal derivation. A grammatical investigation may establish a grammatical conclusion by displaying the relevant uses and the confusion among them. A therapy may establish that a particular question arises from a particular picture by exposing the dependence. But once either says that the question is illegitimate, that no explanation belongs there, or that inquiry should stop, the conclusion still needs to follow from something.
Calling the method non-scientific changes the currency, but the debt remains on the ledger.
If family resemblance describes how the ordinary word “game” is learned, extended, and negotiated, it may show that speakers do not consult a necessary-and-sufficient definition. It may show that no such definition is represented in their ordinary competence. It may show that the lack of one does not prevent successful use. Those are real results about grammar and practice.
None determines what the objects themselves have in common. A grammar of recognition is not yet an account of the recognized object. A report about how a word is used is not yet a result about everything to which the word is applied. If the therapy remains on its own side of that line, my objection does not touch it, and it cannot touch my question.
There is a limit to how personally I can press that complaint. Wittgenstein died in 1951, while Philosophical Investigations appeared in 1953. He had withdrawn the manuscript from publication, but authorized its posthumous appearance. The passages at issue belong to Part I, whose 1945-46 typescript came close to a finished work, so they are not merely a pile of notes assembled by strangers. Yet he never saw the printed book, the disputes it generated, or the portable versions of his name pasted into arguments decades later.
Some line items still belong to him: the absence of one common thing, the disappearance of explanation, the descriptive limit of philosophy, and the promised disappearance of its problems. The text does not discharge every burden those formulations create. Later readers own whatever they add after removing the fine print, but the original bill was already substantial without forging his signature onto the rest.
The interpreters have not found one stable uniform for these sentences either. Some find theses, others therapy, some a positive philosophy, and others an interventionist rather than quietist Wittgenstein. That disagreement has lasted long enough to show that the bare wording does not settle the authority of its own metaphilosophy.
I now knew the dispute was not description against science, but jurisdiction, so I left the word “craving” sitting there for a while to see what happened to it after it left the book.
Then the passport stopped matching the jurisdiction
Once the question about ordinary use was separated from the question about what might actually be common, the troublesome step became almost embarrassingly small.
The text now offered a rather clean bargain. Read modestly, it had already earned its place: ordinary competence does not wait for a complete definition, blurred boundaries need not ruin a concept, and description can loosen the grip of one bad picture. Family resemblance then recommends where to look and what not to assume in advance. It does not decide what the objects share, forbid explanation elsewhere, or turn every search for commonality into a symptom. Several grander conclusions become smaller, but everything left is useful and paid for.
The trouble is that “philosophy only describes” never settles what kind of sentence it is. As a description of what philosophers actually do, it needs evidence and collides with an impressive amount of explanation already sitting on the shelves. As a stipulation for what Wittgenstein calls philosophy, it is coherent and local. As a recommendation, it needs a goal and a reason this method serves that goal better. As a claim about the nature of philosophy, it owes the metaphilosophy required to draw that boundary. As therapy performed on the reader, it may demonstrate a cure in one case without acquiring the medical license for every other question.
The ambition is not merely supplied by later readers. Wittgenstein himself says there is no one thing common that explains the shared word, that explanation must disappear, that philosophy can only describe, and that its problems should completely vanish. A modest reading can narrow those sentences to the grammatical troubles being treated, but the narrowing is an interpretation of language that already reaches farther. Taken with their written force, the remarks have enough authority to close a road and inherit every burden along the route. The safer reading loses the prohibition, while the prohibitive wording loses the safety.
That is the mismatch I could not get rid of. The method enters as one measuring rod among many, carrying one order for one purpose and promising not to make reality conform. A few pages earlier and later, philosophy may not theorize, explanation must disappear, everything relevant lies open to view, and the problems are expected to vanish. The jurisdiction is much larger than the undogmatic passport used to cross the border.
Familiar games look extremely different. Familiar surface properties fail one after another. Then comes the modal phrase: perhaps there is no common essence. Soon the phrase acquires a methodological consequence: stop looking for one common thing, stop demanding a definition, stop treating generality as a possible form of understanding.
That is a great deal of work for perhaps.
The first overreach arrives even before essence, inside the phrase “nothing common”. Common in what sense, at what level, and under which relation? The word is left untyped, while the inspection proceeds through a short list of familiar properties and returns a verdict over the entire space of possible answers.
Looking across a collection can defeat every candidate actually tested, since a counterexample removes the candidate it contradicts. Repeating that operation only expands the list of failed candidates and does not quantify over every candidate that could exist. Amusement, competition, skill, luck, winning, and losing may all fail without taking relations, conventions, abstractions, or higher-order properties with them. Variety in one inventory is not an impossibility theorem over every kind of commonality.
There are two ways to read “there may be no common essence”. If it states only a bare logical possibility, then of course there may not be one. There may also be one. Listing both possibilities establishes no ordering between them and licenses no rule for inquiry. If the sentence is offered as a conclusion supported by the survey of games, it owes the missing inference from failed candidates to the absence of an essence.
The sentence is harmless under the first reading and consequential under the second. It cannot borrow the safety of the first whenever challenged, then recover the authority of the second whenever somebody tries to continue the investigation. No conclusion about essence arrives on credit.
Perhaps enters the room as a hedge and leaves carrying a court order, without anyone having checked its pockets along the way.
There is a tempting escape from this bargain: refuse the burden itself. Perhaps philosophy is therapy rather than theory, perhaps asking for proof merely repeats the confusion, and perhaps the method vindicates itself when the questioner stops feeling the need to ask. But that is not a third reading. It is the ambitious reading keeping its authority while returning the invoice.
And over that return hangs one very large question: why? Why does this diagnosis apply to this inquiry? Why should a result about use govern what the objects share? Why does one successful therapy establish exclusive jurisdiction? Why is the demand for reasons itself another symptom? The method may answer in its own currency, but it cannot declare the cashier a metaphysical confusion and quietly take the goods home.
I also realized that I had been letting several claims share one coat.
Something common may be a feature, a property, a relation, a convention, an abstraction, a higher-order organization, or an invariant under a declared way of comparing things. A necessary condition says every member must satisfy it. An essence does more: it explains what makes each member the kind of thing it is. A complete definition decides membership in both directions. These are not interchangeable prizes sitting at the end of one failed scavenger hunt.
The premise about commonality is the lowest floor. Any common essence must contain or entail something common, but something common need not be an essence, a complete definition, or even the sort of first-order property Wittgenstein’s list examines. Rejecting essence therefore cannot reject commonality in all its weaker forms.
There is a small embarrassment here: if family resemblance characterizes every member through one specified network of overlapping similarities, belonging to that network is itself a common higher-order condition. It may be conventional, relational, trivial, or useless as an essence, but it remains common. If no shared relation is specified, the account loses the means to say which family is being described, and either way, “nothing common” has spent more than the metaphor earned.
Whoever argues against essence need not eliminate every weaker commonality first, but the argument must actually reach essence. Several failed visible traits do not, nor can that unpaid claim veto a search through properties, relations, conventions, abstractions, or anything else the original inspection never reached.
Nor does any of this make essence exist by default. Its existence carries one burden and its absence another. I have no interest in balancing two camps because only claims have standing here. There is no default essence, no default absence, and no victory for choosing the more fashionable uncertainty.
The real irritation was never family resemblance, blurred concepts, or the denial of essence, but premature surrender: allowing a useful warning about one picture of meaning to close an inquiry it has not entered.
So I went back to where this started
The Wittgensteinian approach enters game studies in a respectable form. Jonne Arjoranta treats definitions as tools inside particular discourses, where one purpose may draw a boundary that another purpose has no reason to keep. A field can study communities, interfaces, labor, fiction, markets, or culture without solving the ontology of Game first. “You do not need this definition for your present work” is often excellent advice.
Purpose determines usefulness, though, not truth. A boundary chosen for one study cannot answer a question about the object merely because the study had no use for the answer. “Nothing common need exist” is another claim, and “the search for anything common is a philosophical disease” is stronger again. Practical indifference cannot pay either bill. M. W. Rowe saw why the example matters: once Game carries a wider argument against definition, it stops being disposable evidence.
Then there are the smaller arguments that sent me back to Wittgenstein. Someone proposes a definition, reads one in a game-design book, or asks where videogames end and toys begin. Family resemblance appears in the replies with remarkable regularity.
Sometimes it helps by reminding an OP that ordinary words are learned through repeated encounters rather than necessary and sufficient conditions. Sometimes it attacks the candidate by checking what its criterion includes and excludes. In both cases, the reply contributes something the candidate can answer.
Sometimes a dead philosopher has instead been compressed into a portable reaction image, with the footnotes thoughtfully removed. Wittgenstein already dealt with this. People already tried. Philosophy already moved on. An objection has existed for decades, so apparently it won decades ago, and truth belongs to whoever entered the queue first.
The fine print disappears along the way. Ordinary use becomes the nature of the object, one failed definition becomes the impossibility of definition, and philosophical therapy quietly becomes an ontological result. The separate threads eventually began to read like one patient conversation:
OP: I think the things called games share B.
Reply: Games have family resemblance. There may be nothing common to all of them.
OP: Does family resemblance make B false?
Reply: It means games do not need one essence.
OP: B need not be an essence. It could be a property, relation, convention, abstraction, or necessary condition.
Reply: Then it does not define everything people call games.
OP: I did not say it did.
Reply: Why are you gatekeeping games?
OP: Does that make B false?
Reply: This obsession with definitions is exactly the problem.
Nothing has touched B. The objection began with commonality, retreated to essence, moved to ordinary extension, borrowed moral urgency from gatekeeping, then diagnosed the act of continuing. The candidate stayed in one place, while everything around it enjoyed the walk.
The amount of work proves nothing, and one clean counterexample can erase months of formalization, as it should. The problem is a reply that supplies no counterexample, attacks no premise, identifies no invalid inference, and corrects no scope. The veto can still claim every outcome: a counterexample defeats the candidate, a revision proves the search endless, and a narrower scope is dismissed as incomplete. The candidate can lose while the refusal never states what losing would mean for itself.
I did not conduct a census of Reddit, thank God. I only needed to know whether this creature was imaginary, and it was easy to find across game design, gaming discussion, and philosophy forums. Family resemblance can sharpen a question, but its pocket form prevents the question from surviving long enough to become precise. Later readers did not invent the reach from nothing, though. The gap already existed between Wittgenstein’s measuring rods and his prohibitions.
Gatekeeping gives the veto moral urgency. The suspicion has a history: “not a real game” has demoted walking simulators, visual novels, mobile games, their audiences, and sometimes their makers. A predicate about an object quietly becomes a ranking, an identity test, and an institutional boundary. That history explains the hostility, but it cannot make a premise false before anyone checks it.
The social chain usually runs without stopping: not a game, therefore a lesser work, therefore not made for real gamers, therefore made by someone in the wrong profession. Every arrow deserves suspicion. Yet because those implications are bad, the first sentence is often treated as unavailable before its predicate has been examined. A sequence of invalid implications becomes armor for its own premise.
A classification does not logically contain the status later attached to it. If a community, platform, archive, market, or award converts gamehood into access, that institutional rule performs the gatekeeping. This does not excuse a speaker who used taxonomy as camouflage for contempt, but motive and harm still cannot decide whether the stated condition is true.
The accusation becomes useful when it locates the missing step. It may show that evaluation was smuggled into classification, that the declared domain already contains a counterexample, or that an institution attached punishment without justification. Merely saying “gatekeeping” finds none of these. The word arrives wearing an ethics badge, issues a warrant in logic, and waits for the candidate to apologize.
The debts remain separate. An analyst owes the candidate’s truth within its scope. An institution owes the rule converting classification into consequence. A critic contributes by showing where either debt went unpaid. The label alone pays nothing.
And we are back at the diagnosis
After all that, I returned to the word “craving”.
Wittgenstein is not merely saying that generalization sometimes fails. He connects the desire for it to characteristic philosophical confusions, the seduction of scientific method, and contempt for the particular case. It is a diagnosis of the questioner’s intellectual posture. The question feels compulsory because a picture holds the questioner captive.
Sometimes that is exactly right. A demand for one feature can be a refusal to look at actual differences. An elegant generalization can erase the structure under investigation. A philosopher can spend years repairing a definition whose only purpose is to preserve the expectation that there must be one.
But a diagnosis needs criteria. Otherwise every failed inquiry becomes a symptom after the fact, and every unfinished one can be diagnosed before the result arrives. Continued inquiry proves the craving while surrender confirms the cure, leaving a clinic that no longer needs examinations. Refusing treatment is another symptom, and getting better means agreeing that you were ill.
The text assigns philosophy a therapeutic job, then enlarges it through modal force: theory is barred, explanation disappears, description stands alone, and philosophy can do no more. Restricting those sentences to grammatical troubles may save them, but the restriction does not appear with equal clarity in every claim. A method’s purpose cannot prove the jurisdiction its author gives it.
The method can describe itself and sometimes does so with more caution than its descendants. What remains unclear is whether it can distinguish the limit of its sight from the limit of philosophy. Without that difference, every method becomes sovereign by translating the rest of the world into its own language. Science can reduce literature to cognition, literature can return the favor by making science another story, and linguistic therapy can turn both into uses of words. Redescription is cheap enough for everyone, while exhaustion still needs paying for.
So who is being dogmatic here: the one who denies the question its legitimacy without carrying the burden of that denial, or the one who puts forward a candidate and leaves it open to attack?
What I reject is the strange privilege granted to surrender. An inquiry is required to justify its legitimacy before its candidate can be tested, while the declaration that the question is confused may pass as modesty without proving either the absence or the confusion. Modesty would have left the question open.
I do not need an essence to exist. I am not even convinced that human beings have one, and they matter rather more to me than the dignity of a definition. I do not need every useful concept to possess a sharp ordinary-language boundary. I do not need philosophy to imitate physics, or every inquiry to terminate in a formula. I only need each claim to pay for the distance it travels.
Family resemblance can describe a practice, break the spell of one bad picture of meaning, and remind us that blurred boundaries do not make a word useless. Those are substantial achievements. Wittgenstein’s stronger formulations may travel farther after paying for the distance, but if the text returns the invoice, the last word is not family resemblance. It is the question still standing behind it: why?
What I was reading
- Ludwig Wittgenstein, Philosophical Investigations, especially sections 65-88 and 89-133.
- Ludwig Wittgenstein, The Blue Book, especially the discussion of the “craving for generality”.
- Heather J. Gert, “Family Resemblances and Criteria”.
- M. W. Rowe, “The Definition of ‘Game’“.
- Jonne Arjoranta, “Game Definitions: A Wittgensteinian Approach”.
- Christian Erbacher, “Editorial Approaches to Wittgenstein’s Nachlass: Towards a Historical Appreciation”.
- Alois Pichler, “Outline of an Argument for a ‘Therapeutic’ Reading of Wittgenstein’s Philosophical Investigations”.
- Paul Horwich, Wittgenstein’s Metaphilosophy.
- Oleg Pukhaev, “Understanding Wittgenstein’s Positive Philosophy through Language-Games: Giving Philosophy Peace”.
- Daniele Moyal-Sharrock, “The Myth of the Quietist Wittgenstein”.
- Reddit discussions about family resemblance, repeated attempts to define games, modern definitions of game, sandboxes versus toys, the claim that games cannot be defined, “real” games, walking simulators, visual novels, and gatekeeping gamer identity.
FM39hz, 05/09/2026