Are Questions Just Looking for Answers, or Do They Shape Our Thinking?

The Missing Shoe and the Power of a Question

You wake up late and your other shoe has vanished. You call out, “Where’s my left sneaker?” The question feels simple. But what are you actually doing when you ask it? You’re not just making a statement; you’re setting up a hunt. You’ve opened a space of possibilities: under the bed, in the closet, by the door. Philosophers of language have spent nearly a century trying to figure out what a question is, and why it’s not just a sentence with a question mark at the end. Their surprising answer — still being fought over — could change how you think about thinking itself.

The central puzzle is whether questions are really just declarative sentences in disguise. A declarative sentence is one that can be true or false, like “The shoe is under the bed.” Many early theorists, inspired by the logician Gottlob Frege (1848–1925), believed that statements come first, and a question is nothing more than a way of pointing to a missing piece of information. But a growing number of philosophers now argue that asking is not a weaker cousin of telling. Instead, questions may be the more basic building block of reasoning and conversation.

Questions as Missing Statements: The Traditional View

To see why anyone would think questions are secondary, try turning a question into a set of statements. Take “Did the dog eat my shoe?” This seems to stand for two declarative sentences: “The dog ate my shoe” and “The dog did not eat my shoe.” In the language of philosophy, those statements are propositions — the abstract things that declarative sentences express. On this picture, the meaning of a question just is the collection of all propositions that could count as a direct answer. The question itself doesn’t say anything true or false; it only presents options.

Philosophers call this a proposition set approach. Charles Hamblin developed one version in the 1970s: a question like “Who is coming for dinner?” denotes the set of propositions {Paul is coming, Nina is coming, Paul and Nina are coming, …}. If you manage to settle on exactly one true proposition from that set, you’ve answered the question completely. A different version, the partition semantics created by Jeroen Groenendijk and Martin Stokhof in the 1980s, goes further. It says the meaning of a question is a partition — a way of dividing all possible situations into exhaustive, mutually exclusive answers. For “Who is coming?” the partition would split the world into blocks like “only Paul comes,” “only Nina comes,” “both come,” “neither comes,” and nothing overlaps. That sounds neat: a question carves reality into clear compartments.

For simple cases, both views work beautifully. “Where is my shoe?” carves the world into “shoe under bed,” “shoe in closet,” and so on. The problem arises with questions that aren’t so cooperative.

When Statements Don’t Tell the Whole Story

Consider a conditional question: “If Ann comes to the party, will Bill come too?” In a world where Ann does come and Bill comes with her, what is the single, exhaustive true answer? One option is “Ann and Bill are both coming.” Another option is “If Ann comes, Bill comes.” But those two propositions aren’t the same. The second one is true even in worlds where Ann stays home! If you pick that as the exhaustive answer, then the partition now overlaps (the world where Ann doesn’t come fits both “Bill doesn’t come” and “Bill does come” possibilities). That violates the partition’s rules. Yet pre-theoretical intuition can’t decide which option is better. The question itself seems to resist being squeezed into a static collection of exhaustive statements.

Alternative questions create a similar headache. Imagine asking, “Is Ann coming, or Bill?” The answers “Only Ann comes” and “Only Bill comes” straightforwardly resolve the issue. But what about “Neither comes” or “Both come”? Those feel different — they dismiss the question’s premise rather than choosing between the offered alternatives. A simple partition treats all four answers as equally valid blocks, even though ordinary speakers sense a difference in status. That fine-grained difference is lost.

These puzzles pointed to a deeper flaw: if a question’s meaning is just a set of predetermined answers, then we need a crisp rule for what counts as an answer. The tidy rule — “the true exhaustive answer” — breaks down when the question itself leaves the boundaries fuzzy. An even older voice, the philosopher R.G. Collingwood (1889–1943), had warned as early as 1939 that assertoric thinking — the habit of putting statements first — might be getting logic backward. Perhaps, he suggested, questions are not built out of propositions at all. For decades his suggestion gathered dust, until a new framework gave it teeth.

The Rise of Inquisitive Semantics: Questions as Issues

Inquisitive semantics, developed most fully in the 21st century by Ivano Ciardelli, Jeroen Groenendijk, and Floris Roelofsen, tries to capture what was right about both sides. It agrees with the old proposition set idea that a question presents possibilities. But it adds a crucial twist: the set of possible answers is downward closed. If a proposition α counts as a resolving answer, then any stronger proposition β (one that implies α and gives even more detail) also counts. After all, if “Paul is coming” answers “Who is coming?”, then “Only Paul and Nina are coming” also answers it — you gave too much information, but the question’s issue is still settled.

This simple adjustment has huge consequences. In inquisitive semantics, the meaning of a question is an issue: a requirement to locate the actual world inside one of the highlighted regions of logical space. Statements, too, get a unified treatment. A declarative sentence like “The shoe is under the bed” raises the trivial issue of ruling out only the worlds where the shoe isn’t under the bed. So neither questions nor propositions are more basic; both are moves in a conversation that update what the participants take to be at stake.

The new framework handles conditional and alternative questions more naturally. Instead of forcing every world into a single block of a partition, inquisitive semantics can represent the special status of answers that dismiss the question’s presupposition — they partially resolve the issue but not in the expected way. This aligns with the feeling that “nobody is coming” doesn’t answer “Who is coming?” in quite the same way as “Paul is coming” does. Inquisitive semantics has become a thriving research program, connecting logic, linguistics, and even computer science.

Why “Why” Is a Special Kind of Question

So far, the examples have been whether-questions (“Is X true?”) and which-questions (“Who, what, where?”). But there’s a whole other monster in the question zoo: the why-question. “Why is the sky blue?” isn’t just asking you to pick a statement from a list; it’s asking for an explanation. In the 1960s, Sylvain Bromberger tried to give a formal theory of why-questions using something called abnormic laws — general rules plus lists of special exceptions. But a very different approach, proposed by Bas van Fraassen in 1980, captured something intuitive that Bromberger missed: every why-question hides a contrast.

Van Fraassen argued that “Why is the sky blue?” is usually short for “Why is the sky blue rather than red, green, or yellow?” The hidden contrast class is a set of alternative possibilities. An answer that explains the blueness but doesn’t explain why it isn’t another color would feel incomplete in the right context. If you ask “Why did the polar bear go south?” the relevant contrast might be “rather than staying north” or “rather than swimming west.” Change the contrast class, and the correct answer can change completely, even though the topic proposition stays the same.

This view matters because it shows that the relationship between a question and its answer is not just a formal list of propositions; it’s soaked in context and purpose. Some later thinkers, like Bradford Skow, have pushed back, arguing that answers to why-questions are always causes or grounds, not context-dependent contrasts. The debate is still open. What’s clear is that no one theory of elementary questions can easily swallow all flavors of asking.

Why It Still Matters: Questions Are the Engine of Knowing

In everyday life, you talk about knowing-who, knowing-where, and knowing-why all the time. “Does Grandma know where her glasses are?” These are embedded questions — questions tucked inside larger sentences after verbs like know, wonder, and remember. How should we understand them? If knowing-where is just knowing a true statement like “The glasses are on the counter,” then why do we sometimes say someone doesn’t “really” know where they are, even if they believe a true proposition? Perhaps they saw the glasses earlier but now hold a false belief about a different location. Many philosophers today argue that knowledge with a wh-complement is sensitive to false beliefs and isn’t reducible to plain knowing-that.

Philosopher Jonathan Schaffer has pointed out a sharper puzzle: suppose you know that the cat is on the mat. Do you automatically know where the cat is? And do you automatically know what is on the mat? Those seem like three different pieces of knowledge, even though the fact underneath is the same. This forces us to take questions seriously as the way we organize what we know. The distinction between asking and telling is built into the verbs we use: you can wonder who called but you can’t wonder that someone called. Language itself treats questions as a fundamental category.

So the next time you lose your shoe and shout a question into the house, you’re not just missing a fact. You’re raising an issue, poking a hole in the shared map your family keeps, and inviting someone to patch it. Philosophy suggests that this act of opening possibilities — not just filling them — is one of the mind’s deepest tricks.

Think about it

1. If someone says “I know where the treasure is buried” but they’ve only guessed correctly, would you say they really know? Why might a lucky guess feel different from real knowledge?
2. Imagine a robot that can answer every “yes/no” question perfectly but never asks one itself. Would you say the robot understands the world, or is something missing?
3. Can you invent a question whose answer isn’t a statement at all? If so, what does that do to the idea that questions are just sets of possible statements?