I'm struggling to decide whether to jettison use of the word fact, because the definition appears to be not solid enough to support continued usage. What do I mean by that? Look at one "meaning ladder" (taken from Random House via TFD Online) among several on the same page:


  1. something that actually exists: Your fears have no basis in fact.
  2. something known to exist or to have happened.
  3. a truth known by actual experience or observation; something known to be true.
  4. something said to be true or supposed to have happened.
  5. an actual or alleged event or circumstance, as distinguished from its legal effect or consequence.

This definition marches us from something that exists to something that is merely supposed to be true to something that may be "actual or alleged." (And yes, I am aware that dictionaries don't dictate the meanings of words; they record meanings from usages. And the meanings of this word as it is used and recorded in English seem to be antagonistic toward each other.)

What are we to do with all this? Does a fact require the modifier true to be judged genuine? When we preface a statement with "in fact" don't we mean What follows is the truth? The aforementioned dictionary certainly thinks so:

in fact, in truth; really; indeed: They are, in fact, great patriots.

Here fact and truth are equated absolutely. So I'm wondering: how do we distinguish between what is a fact in the sense of absolute truth and what is a fact of a lesser order? Other words can have many shades of meaning, but this one seems somehow like it shouldn't. So if I hear the word fact without hearing true before it, does it even deserve the term?

A cautionary note

I'm not really looking for a discussion of truth in the philosophical sense. The scope of this question is limited to the meaning of a word in English, not the meaning of an absolute concept as rational beings can or should understand it. What I'm really after, as I mentioned in a comment, is whether the adulteration of this particular word renders it, ultimately, meaningless, and therefore something to be avoided.

In conclusion, I offer this quote from Howard K. Zinn, from his Afterword to A People's History of the United States:

But there is no such thing as a pure fact, innocent of interpretation. Behind every fact presented to the world—by a teacher, a writer, anyone—is a judgment. The judgment that has been made is that this fact is important, and that other facts, omitted, are not important

  • 14
    A fact must be true, but then, what is truth?? (You're asking a question that philosophers have puzzled over for millennia.)
    – Hot Licks
    Commented May 28, 2015 at 11:33
  • 2
    Yet you ask "how do we distinguish between what is a fact in the sense of absolute truth and what is a fact of a lesser order?"
    – Hot Licks
    Commented May 28, 2015 at 11:37
  • 3
    @Robusto: Your dilemma can be restated: Can I use a word precisely if other people use it loosely? Definitions found in dictionaries simply describe the word as it is used in various contexts. Any prescriptive authority boils down to "people don't use the word that way in that context". Context is king, also in lexicography.
    – TimR
    Commented May 28, 2015 at 12:49
  • 4
    'a true fact' is a pleonasm, yes. But 'in fact' is an idiom meaning 'it has been established'. Also, people may use 'in fact' wishfully. I don't think this is like 'literally'.
    – Mitch
    Commented May 28, 2015 at 12:52
  • 5
    What resources can we provide? What evidence do you require. It is, I suspect, one of those interesting, but subjective, questions which cannot have a "right" answer. Ultimately you, the OP, will choose the answer which you "agree" with most, and not the most objective one (ironically an answer containing hard cold facts). If this had been asked by anyone else, it would have been immediately closed as POB.
    – Mari-Lou A
    Commented May 28, 2015 at 18:33

15 Answers 15


A fact does, in fact, have to be the truth at the time you're using the word.

By 'truth', I mean something you believe to be true (due to any of several possible reasons).¹

Consider: "The number of planets in the solar system is eight." A few years ago, this was not a fact. It is now. (Just an example, don't attack the example.)

[EDIT: Before 1917, people thought it to be a fact that the atom was the smallest particle of matter. Today, it is a fact that it isn't, and we were wrong in thinking it was... I'm including this example to show that in light of the new fact, what we thought was a fact previously, can be rendered untrue for that time too; something @Jay pointed out my previous example didn't specify]

When a jury convicts a man, it's a fact that he's guilty. If he's later acquitted, it's a fact that he is 'not guilty'. (As far as the public is concerned. Individuals who actually saw the crime might know, for a fact, whether it's true or not)

What I'm trying to say is that the word fact is used for what you know (or sincerely believe) to be the truth or what is widely believed to be the truth at the time of speaking. Facts are subject to change.

Something is not a fact if you know/believe it to be untrue or if it can be easily be shown to not be widely believed at the time.

The two can contradict. That's when myth comes into play.

Tom: "Interesting fact: you can see the Great Wall of China from space." (Widely believed).

Neil: "As a matter of fact, you can't. That's a myth." (I went to space. I know better.)

¹ Please note that I'm talking exclusively about the cases where you use the word fact. In those cases, I infer you strongly believe it to be true.

  • 11
    Robusto: Tom's usage in "Interesting fact" is not a slippage of meaning of 'fact', it is just wrong. Just as if you point at a dog and say 'cat'. You may mean 'cat' sincerely or have bad eyesight or have had an aneurysm, but either way it is a mistake.
    – Mitch
    Commented May 28, 2015 at 13:36
  • 3
    The phrase "a fact in law" is not uncommon, IME, and reflects the idea that the legal meaning of guilt and innocence is not, in fact, definitive and a legal fact may really be a fiction (a "technicality") or even nonsense ('"for the law supposes that your wife acts under your direction." "If the law supposes that," said Mr. Bumble, squeezing his hat emphatically in both hands, "the law is a ass..."' — Dickens, Oliver Twist)
    – Nagora
    Commented May 28, 2015 at 14:13
  • 2
    I (a lawyer) have never heard the phrase "a fact in law" in legal practice. Fictions are not facts, nor are presumptions. They are things treated as true even if either they aren't or are not known to be. I doubt whether there was ever a presumption that a wife acted under husband's direction so as to assign the wife's criminal actions to the husband. I don't think that is how couverture worked. Commented May 28, 2015 at 15:03
  • 1
    @sgroves: You're right. But the example works better with moon. And Neil. Neil Armstrong did actually go on the record debunking this myth.
    – Tushar Raj
    Commented May 28, 2015 at 17:34
  • 3
    "When a jury convicts a man, it's a fact that he's guilty." The takes that as a fact, but it is not a fact. "If he's later acquitted, it's a fact that he is 'not guilty'." Even the law does not take that as a fact. Commented Nov 11, 2020 at 18:46

There are some good answers here. I thought it would be useful to expand on meaning (5) in your list, in case you ever have to deal with it.

In English common law, most matters (criminal or civil) came to be tried by a jury. Over time a rule developed that some questions would be decided by the jury and others by the judge. The jury decided questions of fact whereas the judge decided questions of law.

For example, in the English law of theft, one element of the crime that must be established by the prosecution is that the defendant acted "dishonestly". The meaning of "dishonest" is a question of law, but whether the defendant was dishonest according to that definition is a question of fact.

There are practical consequences of this distinction. In English criminal procedure, you may appeal the judge's decision on questions of law, but not the jury's decision on questions of fact.

In this usage "fact" is used in a slightly technical sense in opposition to the word "law" but it is essentially intended to be restricted to statements that are true, at least that is the hope.

  • 1
    +1 Thanks for that clarification on meaning 5, I knew it had to me something like that. And apologies for the cheap shot in my answer...it just seemed to fit the argument I was making too well.
    – JeffSahol
    Commented May 28, 2015 at 17:53
  • @JeffSahol - no worries. I both practice and teach law and I am very well used to the cheap shots. Commented May 28, 2015 at 19:02
  • 1
    Just want to add that this is in fact the historical definition of the word, all the way up to the 16th century or so, and hence its persistence in the English legal system.
    – Yang
    Commented May 29, 2015 at 7:13
  • 1
    The word "alleged" used in the dictionary should also be seen in a legal context, e.g. "al-legation", to support a plea/legal claim, and under law, of course, all allegations need to be scrutinized. This is different from the common use in which use of the qualifier "alleged" is often inherently mixed with a heavy dose of skepticism.
    – Yang
    Commented May 29, 2015 at 7:16

Your question sounds as if you are in a Quixotic fight against a particular false Scotsman, not realising that there is no such thing as a true Scotsman, or for that matter, a true X for any word X whatsoever.

You could just as well be "struggling to decide whether to jettison use of the word word, because the definition appears to be not solid enough to support continued usage". There is no clear boundary between what is a word and what isn't. Is "ouch" a word? Or "hmmm"? How about "Aargh"? "Aaaaaaaaaaaaaaaaargh"? Is "blasé" an English word? What about "révisioniste"? Is "Agatha"? And "Christie"? Is the symbol used by 'the artist formerly known as Prince' a word? Is "a priori" an English word? Or is it two? How about the two constituents? Note also that in languages which do not separate words by spaces, it can be practically impossible to distinguish separate little function words from suffixes or prefixes.

Who is this I who is struggling anyway? Do you mean your body, your mind (whatever that is) or your soul (whatever that is)? Did this I already exist when 'you' were born and possessed only a tiny percentage of the atoms that you would now consider your own, whereas many of the atoms 'you' possessed then have found something else to do in the meantime? Did it already exist before you could think clearly and create memories that you can still draw on?

What do you mean by struggle to decide? Who is this ominous person trying to prevent you from deciding, the one you apparently have to struggle against?

What do you mean by the definition? Since when do words have unique, well defined definitions? Where would you be able to find the one true definition of a word? How would you be able to understand it without using similar definitions for all the words used in it, leading to loops and an infinite regress?

Facts are true because (in some sense that I don't wish to make precise) the meaning cloud for the word fact is mostly a subset of the meaning cloud for the word true. Although false facts are mostly outside the one for true, while still in the periphery of the one for fact.

For similar reasons, birds can fly even though some birds have broken wings or happen to be penguins. And facts are true even though some are not.

  • 2
    Nicely put: "For similar reasons, birds can fly even though some birds have broken wings or happen to be penguins. And facts are true even though some are not."
    – DCShannon
    Commented May 28, 2015 at 21:26
  • This answer... is beautiful.
    – wavemode
    Commented May 29, 2015 at 2:15
  • 4
    Except the poster said he was NOT asking for a philosophical discussion on the nature of truth, but simply a clarification on the definition of the word "fact".
    – Jay
    Commented May 29, 2015 at 6:08
  • 1
    @Jay: True enough, yet I enjoyed this answer so much that I voted it up. :)
    – Robusto
    Commented May 29, 2015 at 11:47
  • @Hans: The person I'm struggling with to decide is, of course, myself.
    – Robusto
    Commented May 29, 2015 at 20:11

Does a “fact” have to be true?


Here is a detailed definition of fact from OED for the sense that we are dealing with:

A thing that has really occurred or is actually the case; a thing certainly known to be a real occurrence or to represent the truth. Hence: a particular truth known by actual observation or authentic testimony, as opposed to an inference, a conjecture, or a fiction; a datum of experience, as distinguished from the conclusions that may be based on it.

But it is also mentioned that:

Where the truth of a matter is disputed or in doubt, this sense overlaps with sense "A piece of information allegedly or conceivably true; something presented as a fact but which is disputed or unproven; (more strongly) an unproved assertion, an allegation."

In fact, the truth is, it depends.

You understand that the fact is not a truth from the context. Usually, the negative words like false, unproven, disputed etc. reveal that. (They might modify fact also.)

Some examples from OED:

  • This is..a false fact, supported by a supposed motive - 1824, Westm. Rev.
  • It bases its accusations on false statements and inaccurate facts. - 1941, A. M. Lindbergh Diary

  • Waksal hotly disputed some of the facts in that story. - 2002, Vanity Fair

Note: Of course it can be discussed or interpreted further but I focused on the usage of the word.

  • 2
    Another example occurs in the expression "You've got your facts wrong," which, as I understand it, means "The instances or data points you are putting forward or are basing your argument on are not valid." The interesting thing (to me) about this expression is that it doesn't say, as it could have, "Your facts aren't facts [that is, factual]." Instead, it emphasizes that the instances or data points are not valid, even while conceding that they may still be called "facts."
    – Sven Yargs
    Commented May 28, 2015 at 21:50
  • @SvenYargs: Good catch. I prefer saying "Those aren't the facts" myself.
    – Tushar Raj
    Commented May 29, 2015 at 3:43
  • To me the caveat reads the same as if, considering the meaning of the word "horse", the dictionary said, "where the species of a creature is disputed or in doubt, this sense overlaps with the sense, "a creature allegedly or conceivably a horse: something presented as a horse but which is disputed or unproven"". Normally I wouldn't say "your facts aren't true, and I'll prove it", and I wouldn't say "your horse is a cow, and I'll prove it". But you certainly can, in both cases using "X" to mean "so-called X" or "alleged X". Am I missing a subtlety that applies to facts but not horses? Commented May 29, 2015 at 13:39
  • ... or is this just a particularly noteworthy case of the general phenomenon for any word, that it can be used provisionally when the truth is in doubt? Commented May 29, 2015 at 13:41
  • "Where the truth of a matter is disputed or in doubt", I believe you're confusing this in relation to the legal term "question of fact". I don't have a copy of OED to confirm this, but I would guess that's the case. Commented Nov 11, 2020 at 18:59

I think that you are reading too much into the last two definitions listed. The fifth one is apparently a legal term of art and as we all know lawyers have little use for the actual truth. The fourth covers instances where facts are asserted without being verified. If I asked a group of toddlers for a list of facts about where babies come from, that's still a list of facts.

Finally, the same argument can be made for truth, "a fact or belief that is accepted as true" being one of its meanings. So now we're trespassing into the realm of epistemology.stackexchange.com.

  • 'Instances where facts (let's say sense 1) are asserted as such without being verified' are extremely common (have a look at a few answers here). If RH labels 'statements asserted as facts' as 'facts', then this forms the hub of the question. It cannot reasonably just be dismissed. Commented Nov 11, 2020 at 17:42

"What I'm really after... is whether the adulteration of this particular word renders it...meaningless, and therefore something to be avoided."

The word is not being "adulterated"; it is being used in different ways in different speech contexts.

A fact is that which exists or existed, happens or happened, and which can, as a result, be *known, thought, supposed, believed, stated, averred, alleged, etc".

These ancillary definitions involving speech contexts are not really definitions of "fact" per se but of the kinds of mental attitudes we can express, or the kinds of predications we can make, in respect to ontological fact. The lexicographers have made a leap from ontology to (everyday) epistemology.

  • You say the word "is being used in different ways in different speech contexts." That's fine for most words, but that's part of what I'm trying to establish. When is it OK for a "fact" not to be true?
    – Robusto
    Commented May 28, 2015 at 19:05
  • 3
    An ontological fact exists; it is neither true nor false. "True" is an attribute of predication. A statement is "a statement of fact" (i.e. true) when it comports with the ontological fact(s) it's about. A statement that does not comport with the ontological fact(s) is a falsehood. When someone knowingly calls a falsehood a "fact" (by which they mean "a statement of fact"), they're said to be lying.
    – TimR
    Commented May 28, 2015 at 19:34
  • @TimRomano: Beautifully put. May I suggest editing that into your answer. +1.
    – Tushar Raj
    Commented May 29, 2015 at 3:48

When I refer to something as a fact, without qualifications, I am implying that it is true. This situation is not much different than when I say something like "I ate a grapefruit for breakfast." I am implying that I actually did. I might be lying, but the meaning of my sentence is that my eating of a grapefruit is the truth. I can refer to an event as alleged, no longer implying it is necessarily true: "Allegedly, he ate a grapefruit for breakfast." Similarly, I can use the word fact to refer to that event, whether true or not: "The alleged fact of your eating a grapefruit is immaterial to the question at hand." It is possible to drop the word alleged in this last sentence, and without it, the sentence will be ambiguous about whether I am implying that you did eat a grapefruit. Some people will interpret it to mean I implied you did, others will interpret it to mean I am agnostic.


As the definitions you quote indicate, the word "fact" is used in two different senses. It can mean something that is true, or it can mean something that is claimed to be true, but which may or may not actually be true.

On the one hand, people will say things like, "It is a fact that Senator Jones accepted a bribe." Meaning, this statement is true. Or conversely, "No, that's not a fact at all." Meaning, that statement is false.

On the other hand, people also commonly say things like, "The facts are in dispute", meaning, we are arguing about what is true and what is not. Editors talk about "fact-checking" a document to determine whether the facts it states are true or not. We talk about "unproven facts" or "questionable facts". We say, "You have your facts wrong." Etc. If the word "fact" was understood to mean "statements that are true", then it would be a paradox to say "unproven facts" or "disputed facts" or "the facts are wrong". How can something that we all know is true be in dispute, etc?

  • A statement can be both true and unproven or disputed. Commented Nov 11, 2020 at 18:50
  • @Acccumulation True. But if it's unproven or disputed, than by definition we don't know whether it's true or not. If the word "fact" means that something is true, than to refer to a statement as a "disputed fact" would be somewhere between a contradiction -- is it disputed or is it a fact? -- and an assertion that it is true but some people continue to deny it.
    – Jay
    Commented Nov 12, 2020 at 14:08

From Hard Facts: nuances in meaning and usage exist, but there are good reasons to use both terms:

  • What is the clear difference between a fact and a truth? Well, if you look into most dictionaries, you will be amazed to find that the two words are actually very close in terms of their definitions. This is because the two terms are very much related. That’s why you really can’t blame people for recognizing both as similar terms.

  • Fact is basically something that exists, or is present in reality. Hence, these are things that can be seen visually, and these are the things that can actually be verified. Facts are objective matters rather than subjective ones. It is not just something that you believe, but rather these are more or less the things that can be observed empirically, or by the senses. So, facts can be seen and heard, as well as proven by the other senses.

  • Truth can be described as the true state of a certain matter, may it be a person, a place, a thing or an event. It is what a person has come to believe. If he believes that something is true, then it is true. It also answers the questions of what’s really happening. In the technical sense, facts can answer certain ‘why’ questions, like ‘where’ or ‘when’, and even ‘how’, while truth answers the question ‘why’. The question of ‘how’, and even ‘what’, are said to be answerable by either of the two.

  • In terms of permanence, a fact happens to be more permanent, and almost always seems to have no changes. It is more constant than truths. For example, when you say that the sun will always rise from the east and set in the west, you are telling a fact, but when you say that you are in Los Angeles, then that is a truth, at least for that exact moment. Several hours from that time you may have gone somewhere else, making your previous statement a fallacy. Thus, a truth is something that is not universal, it is more subjective, and depends on the current situation. That’s why the truth’s existence is said to be more temporary than that of facts.

    1. Facts are more objective when compared to the more subjective truths.
    2. Facts are more permanent when compared to the more temporary truths.
    3. Facts exist in reality, whereas truths are usually the things that one believes to be true, or the things that are true in the current situation.
    4. Facts can also answer the ‘where,’ ‘when’ and ‘how’ questions, whereas truths answer the ‘why’ question.

If you want to know how the word "fact" is used in English, that is quite different from what philosophers or mathematicians consider "fact". Here are some more definitions like the one you provided which seemed to prompt you to ask this question:

c. Something believed to be true or real:
American Heritage Dictionary

2:a piece of information presented as having objective reality
("presented as having" does not refer to "real" fact - whatever that means)
Merriam-Webster Dictionary

Other definitions are along the lines of:

"information known to be true".

Would you admit that I can "know" something and be wrong? If that's true, then all other dictionaries allow for "fact" to mean something which is NOT true scientifically/objectively/verifiably.

So if you are interested in how the word "fact" is used in the language itself, forgetting the rigorous definitions of truth and fact that have to do with verification, and leaving apart the special study of epistemology and all the philosophical stuff, it's clear to me that when used in our language "fact" does not always mean something that is true.

Let's say the basis of whether something is fact is a definition, like the IAU's definition of a planet. Pluto is no longer a planet, but a dwarf planet, making the statement "Pluto is not a planet" a fact. Suppose the IAU change the definition tomorrow. Has the fact changed?

  • So many people making the same mistakes. It's a fact that the IAU's previous definition of a planet included Pluto. It's a fact that now it doesn't. It's a fact that there now considered to be eight planets in the solar system. These are indisputable statements, which means they are facts. People keep using false statements as if they were facts in order to demonstrate that facts are false. It's maddening. Commented Mar 23, 2020 at 22:40

There are basically two types of facts: Universals, which are universally true and is usually part of common sense about how the physical world works, and Facts that are statements you believe are true, but not necessarily.

In your precise case, about the usage of the word "fact" for communication purpose, we must also account for the incompleteness and incertainty of the interlocuter: he/she may be using a prejudice that is just false, or be lying, or just misestimating his/her level of confidence in the statement he/she made.

So, in the end, a fact can only be what one believe is true, as opposed to what is true universally.

BTW, this is a very important topic of research in knowledge representation, logics, artificial intelligence and in any epistemological system (ie, systems that represent knowledge) in general.

  • 1
    ontology is being, epistemology is knowing.
    – TimR
    Commented May 28, 2015 at 19:40
  • 1
    Your argument sounds like we can't REALLY know anything. I agree, technically.
    – Zebrafish
    Commented Jan 19, 2019 at 4:41
  • 1
    See, here we have one particular metaphysical system being asserted: the medieval view that truth is about universals, while facts are contingient statements about Aristotelian accidental properties. And the person making it isn't aware that this is an entirely different conception of truth and fact than a modern scientist would make, using a metaphysics which denies the existence of universals; or that a post-modernist would make, defining "facts" as social conventions; or that a pragmatist would make, defining "truth" as a praxis which proves useful.
    – Phil Goetz
    Commented Feb 5, 2019 at 2:58
  • 2
    @Zebrafish Are you sure? Commented Nov 11, 2020 at 17:45

I'm not really looking for a discussion of truth in the philosophical sense. The scope of this question is limited to the meaning of a word in English, not the meaning of an absolute concept as rational beings can or should understand it.

[I've rewritten this in response to comments.] I'm afraid you are looking for a discussion of truth in the philosophical sense. Nearly everyone agrees that facts are necessarily true; they disagree on what "true" and "necessary" mean. They also disagree over the law of the excluded middle, which says that every claim is either true or false.

(A) Philosophers, Christians, Hegelians (including Marxists, Progressives, and Nazis), humanities professors, and post-modernists believe that "true" means "logically proven to be absolutely, universally, eternally true". This has been the standard philosophical definition of "truth" since Plato. I'll use "True" (capital T) to denote this meaning of truth. They also believe the law of the excluded middle is True (also standard since Plato; made explicit by Aristotle).

Under this system, the claim "Cops are racist" must be either True or False (law of excluded middle). If True, then every cop, everywhere, who ever existed or will someday exist, is/was racist. If False, then no cop anywhere has ever been, nor ever will be, racist. Aristotle would allow for the possibility that some cops are racist and some are not; but he would say that in that case, being or not being racist was an accidental rather than an essential property of cops, and therefore we could know nothing and say nothing about whether cops are racist, since "knowledge" by definition (for rationalist philosophers) refers only to essential properties.

People holding these beliefs use neither quantifiers nor quantities in their arguments, since quantification is, to them, unnecessary. Any claim can be proved True by finding a single instance, or False by finding a single counter-example. (They ignore Aristotle's proviso that some attributes are accidental, because they're not as smart as Aristotle. Anyone as smart as Aristotle wouldn't be an Aristotelian in the 21st century.) This belief system is ineffective at understanding the world, but effective in politics, since it lets you prove anything you want to from just a single example.

(Post-modern philosophy is nothing but the belief that this is the True definition of "true", combined with the observation that nothing in life ever attains this standard of True or False, applied to "deconstruct" whatever the post-modernist doesn't like.)

(B) A typical scientist, engineer, contemporary empirical philosopher, or free-market economist believes that "true" refers not to the human-language claim X (in words), but to a statistical claim about the frequency with which, or the probability that, some quantitative restatement of "X" holds.

Suppose that X = "Artificial food coloring makes children hyperactive".

A scientist must do something like this to "prove" that X is "true":

  1. Specify some repeatable observable phenomenon O. Here O is a time-series comprised of (A) a child eating food with artificial colors, and (B) a child's behavior after doing so.
  2. Specify some operationalized measurement m of the degree to which X is true for a given observable datapoint o. In this case, there's a standard operationalized checklist for behaviors indicating hyperactivity; a child's hyperactivity score m will then be how many times they acted out any behavior on the checklist within, say, 30 minutes of eating a cookie with 1 gram of some kind of food coloring.
  3. Propose 2 mutually-exclusionary hypotheses, X and Y, about how every instance o is generated. In this case, hypothesis X is "eating 1 gram of this food color increases frequency of hyperactive behavior". Hypothesis Y is "hyperactive behavior is uncorrelated with eating 1 gram of this food coloring."
  4. Find Q unbiased samples of N instances of such observational data. In this case, find N "typical" children, who are, in genetics, environmental background, age, weight, sex, health, etc., representative of the population to which your claim X will be applied.
  5. Divide your sample into two groups, G1 (who will eat food coloring) and G0 (who won't). In this case we will actually just test all of the children at least twice, getting food coloring half the time, and no food coloring half the time.
  6. Restate X in terms of your measurement, as X'. In this case, X' = "For every child, average hyperactivity scale score will be higher in the half hour after eating a cookie containing 1g of (the specific food coloring)."
  7. Run the experiment, and record the hyperactivity scores and compute the mean and variance for group G1 (tests after eating food coloring) and G0 (tests after not eating food coloring).
  8. Let ni = number of observations in group i. Compute the sample mean mi and sample variance vi of the hyperactivity scores of groups G1 and G0: m1, v1, m0, v0.
  9. Perform a T-test, or an F-test, or some other standard statistical test, at the 95% confidence level. That means you find the value t of a function T(m1,v1,n1,m0,v0,n0) such that the probability P that {a set of N random observations generated by hypothesis Y would, by chance, have T(m1,v1,n1,m0,v0,n0) >= t} is 0.95. In the food coloring example, we'd use a between-groups t-test.
  10. If T > t, we say that, if our samples were unbiased, the claim X' (not X) has been proven with confidence 0.95. You could say X has a truth of 0.95. Note that it's theoretically impossible to ever prove something is True, meaning it has a truth of 1.0.
  11. If instead P > 0.05, we do NOT say we've disproven X'. We say that we failed to prove X' at the 95% confidence level.

(The claim X', that hyperactivity score of all children goes up after eating cookies, failed repeatedly to be proven at the 95% confidence level. Unfortunately, every medical journal article reporting such an experiment incorrectly stated in its conclusion that it had disproven the hypothesis that food coloring increases hyperactivity. As I just explained, that conclusion is not licensed. Decades later, the claim X, that eating food coloring causes hyperactivity, was proven to be true in children with particular genetics. The claim that failed to be proven at the 95% confidence level was that food coloring causes hyperactivity in all children. Furthermore, it could still be true that food coloring causes hyperactivity in all children, but by such a small amount that you'd need a much larger sample size to prove it at the 95% confidence level.)

(C) Someone with a better understanding of math and of language would never regard anything stated in a human language as capable of being shown "true" or "false". Rather, they would take such a statement and try to completely operationalize it in terms of primitive features, ideally down to the level of, say, pixel values in an image. Then they would compute the predictive value of the discrimination performed by their operationalized claim over a large, typical (not "unbiased") data set. Those claims which produced the most predictive value (measured in bits of information) per computation (measured in bits of computational entropy) would be retained. They wouldn't be called "true", but "useful". This is what deep learning networks do.

  • Which is why the quoted passage has an ending, one you apparently did not feel compelled to address.
    – Robusto
    Commented Jan 19, 2019 at 2:29
  • Of course you can isolate the meaning of a single word that way. Colloquially acceleration generally means going faster. We hardly ever say the car is accelerating when it's braking. In physics however and even approaching the question philosophically it's still acceleration, just in an opposite direction. It's like saying retreating is advancing in the opposite direction. The point is the question is about the English as it's spoken by ordinary people.
    – Zebrafish
    Commented Jan 19, 2019 at 6:08
  • @Zebrafish Neither you nor Robusto understand my answer. There is no common de dicto "colloquial meaning" used by "ordinary people" for "fact", because there is no common colloquial meaning, even among "ordinary people", to "truth". So you can only give a de re definition of fact, not a single de dicto definition. Look up "de re / de dicto" if you don't know what that means.
    – Phil Goetz
    Commented Feb 4, 2019 at 22:21

An etymological note: fact comes from a term meaning simply "something done", as etymonline says.

fact (n.)

1530s, "action, anything done," especially "evil deed," from Latin factum "an event, occurrence, deed, achievement," in Medieval Latin also "state, condition, circumstance," literally "thing done" (source also of Old French fait, Spanish hecho, Italian fatto), noun use of neuter of factus, past participle of facere "to do" (from PIE root *dhe- "to set, put"). Main modern sense of "thing known to be true" is from 1630s, from notion of "something that has actually occurred."

A similar phenomenon has occured in Spanish, where two important synonyms for the English "fact" are hecho and dato, the latter which is immediately related to "datum" in English (see etymologias.dechile.net). This, because Latin facta is related to when [a letter] was made, and datum is related to when it was given, which, by the way, is another synonym for "fact" in English...

So the sense of a "changeable datum" has departed significantly from it's etymological origin, as indeed many words have. When once you have given something to someone, you cannot change the fact that it is given, even if you take it back; but a piece of information, once separated from its source, can be changed.

Don't jettison the word, but use it with care.

Etymology aside, here's the word, in the wild:

And as Felsenthal noted, it is one that will be put to the test. “I don’t think there’s ever been a president and vice president to take office in a moment like this, where we don’t just disagree on issues,” he said. “We disagree on basic facts.” (Washington Post, 10-12-2020)

  • 1
    A usefull answer to an awkwardly broad question. Commented Apr 4, 2020 at 1:05
  • 1
    The first part of this answer commits the etymological fallacy. The part about Spanish is not relevant to the question.
    – Rosie F
    Commented Apr 16, 2020 at 14:10
  • Rosie F: No part of this answer commits the etymological fallacy. It simply states the etymological origin of the word, and the fact that it meaning has departed from its original meaning. The etymologias.dechile entry is related because it is the source explaining the relationship between "fact" and "data".
    – Conrado
    Commented Aug 20, 2020 at 13:59
  • related: alternative facts
    – Conrado
    Commented Oct 14, 2020 at 20:42
  • 1
    I would replace "note" in the first sentence with "fact."
    – phoog
    Commented Nov 12, 2020 at 0:40

I'm struggling to decide whether to jettison use of the word fact, because the definition appears to be not solid enough to support continued usage.

Very noble. A futile gesture sometimes succeeds. Think of the population of the English speaking word – 20% of 7.8 billion. If a few billion get behind you, you may become part of history.

The beauty of language is that we need not know the precise definition of all the word we speak or write as many of them are subjective and/or contextual.

Definitions do not really exist save in mathematics: the number “four” is understood as being “the sum of two and two. Two has its own definition that is based on the definition of one plus one as expounded by Whitehead and Russell.

Expecting there be one correct definition for a word is unreasonable - for one or two words would be good. What you need is a set of words to explain another, plus some examples in context.

What replaces definitions in language are close approximations that more or less capture the essence of a word within a context. I offer you the word “window”: the image in your brain is not the same as mine – my window has no glass, or it could be a window of opportunity.

If I say “set” (and nothing else) to you it is meaningless – it has 430 entries in OED comprised of 60,000 words and many sub-entries – and you are expecting one true definition of “fact" or, in a fit of pique, you will abandon it, and thus teach the English-speaking world a valuable lesson.

You choose “fact” – I would suggest you may think about abandoning “for” and a load of other prepositions. The word, “word” presents problems in certain contexts – Word!


There are two seemingly contradictory definitions of the word "fact", however there are important distinctions to be made to understand when they're being used.

The first definition states that a fact is a statement that has to be indisputably true.

The ever reliable OED confirms this:

fact n. a thing that is known to be true

As does the NOAD:

fact |fakt| noun a thing that is indisputably the case

And the Collins COBUILD dictionary:

fact - A fact is an item of knowledge or information that is true.

However, there is also another definition, meaning something in dispute. This is acknowledged by the OED with an alternative definition:

A piece of information allegedly or conceivably true; something presented as a fact but which is disputed or unproven

So it is possible to use the word "fact" to mean something believed to be true, however only if it is not disproven.

This is an important distinction, and worth repeating: If a statement of belief is proven false it is not a "fact".

Consider Donald Trump's statement that more people attended his inauguration that any other President's in history. He may believe it with every fibre of his being, but despite his belief, the statement is disputed by photographic evidence:

enter image description here

Other times people may be careless with the word "fact". For example, the accepted answer here uses the following example:

Before 1917, people thought it to be a fact that the atom was the smallest particle of matter. Today, it is a fact that it isn't, and we were wrong in thinking it was...

However before 1917 it would have been careless for someone to write, "it's a fact that the quark is the smallest particle of matter". Typically such statements are written as, "it's a fact that the quark is the smallest known particle of matter". The first statement is a belief only (and open to being easily disproven by later science), the second is an indisputable statement of truth: A fact.

Here are some real examples to demonstrate this is how such statements are phrased in the real world:

Anyone saying, "it's a fact that X is the smallest particle of matter", would be making a statement that could easily be disproven in the future, because how can we even be completely sure we know the smallest particle of matter? It would be very careless and unscientific to use such language.

Even if they did, it would be a very poor argument because there's no evidence that something smaller doesn't exist.

The accepted answer also uses the following example:

When a jury convicts a man, it's a fact that he's guilty. If he's later acquitted, it's a fact that he is 'not guilty'.

Again, both statements would likely never be used in reality as it's very possible a case could be overturned. All anyone can say after a court declares a guilty verdict is, "it's a fact that this person was found guilty in a court of law".

Even a court of law acknowledges its fallibility, which is why the statement given by the jury when giving their verdict is, "We find the defendant guilty" and not "The defendant is guilty".

This continues in all reporting of the trial, too:

Whether the person is actually "guilty" or "not guilty" remains open to be proven false, and so is not used in real life.

However, again, if someone did, they would be making a statement that would be very hard to backup and claim was a "fact".

It would be akin to stating, "it's a fact that there are no aliens in outer space". Impossible to prove true without venturing into every nook and cranny of the universe, however this statement meets the second definition of "fact".

That all said, if someone found an alien in outer space, then this statement would no longer be considered a fact by any definition. Make sense?

Understanding can change over time, too, as mentioned in the accepted answer.

For example, prior to August 2006 it would have been factual to say, "there are nine known planets in the solar system". After Pluto was reclassified as a "dwarf planet", it has since been factually correct to say, "there are eight known planets in the solar system".

Both statements were facts at the time they were made, according to how humans classified planetary bodies when the statements were made. It's when the statements were made that defines if they were facts or not.

In the real world, although a "fact" can be a belief that hasn't been disproved, it's rare to push that usage when it can be avoided... because it's possible you could shown to be wrong.

People tend to be cautious of making statements which can later be disproven, and the greatest chance of something remaining a fact is to start with the indisputable truth.

In short, if a statement has the potential of being false, as it may not remain a "fact" for very long.

  • You're oversimplifying this, I'm afraid. Rightly or wrongly, anything can be called a fact, even if it's not verifiable. Beliefs, opinions, and judgment calls are often confused with facts. Even NOAD seems to be inconsistent with its example: "She lacks political experience" could easily be an erroneous assumption based on perceived stage fright or a bad outing at a debate.
    – J.R.
    Commented May 29, 2015 at 20:43
  • 2
    It bends my brain that an English language community has fallen into the trap of debating philosophy. A fact is a true statement. If it's determined to be untrue, then it's no longer a fact (and indeed, never was a fact). Commented Feb 5, 2019 at 10:12
  • 1
    OED, M-W and AHD provide alternative, looser definitions that it is unscholarly to neglect even to mention. Commented Nov 11, 2020 at 20:51
  • 1
    @EdwinAshworth That is simply untrue. The definitions you mention pertain only to the legal terms "question of fact" and "finding of fact". Even in a legal sense the word "fact" has the same meaning I ascribe here. I could add countless other dictionaries, but that would be pointlessly excessive. And it's worth pointing out that no other answer has listed every other dictionary, and most don't even have any references. One rule for the answer you disagree with, another rule for those you don't... Commented Nov 12, 2020 at 11:33
  • 1
    @EdwinAshworth 1/ I did not adjust my answer to match another. I added the information you flagged as missing. What was the purpose of your comment if not for that? 2/ I've again re-written that section of the answer you flagged. 3/ This answer adds important distinctions and examples which are missing from other answers, including the overly simplistic accepted answer. Commented Mar 1, 2021 at 17:25

Not the answer you're looking for? Browse other questions tagged or ask your own question.