# What is the difference between "not true" and "false"?

What is the difference, if any, between "not true" and "false"?

• Looking besides the crude restrictions of classical logic: Just think of "neither", "both", and especially "we do not know" = "indeterminable". All these can be labelled "not true", but "false" wouldn't be appropriate. Oct 19, 2016 at 23:00
• "This claim is false" is certainly not true. But I don't think you can claim it's false either, unless you define "claim" in an unusual way. Oct 19, 2016 at 23:45
• it seems simple but in fact it's one of the most complex questions in modern logic.
– user20153
Oct 19, 2016 at 23:48
• Questions like this differentiate logicists, who care about meanings, and formalists, who don't care about meanings. Oct 19, 2016 at 23:57
• there are many logics on offer today. some are 2-valued (e.g. true or false) but there are also many-valued logics. there are still others that do not involve truth values. so this is a complex question. in classic 2-valued logic not true is the same as false. but in e.g. intuitionistic logic you cannot go directly from "not true" to "false". Google "law of excluded middle" to see why.
– user20153
Oct 20, 2016 at 19:55

"p is false" implies "p is not true", but not vice verse because p can also be nonsense.

"2 + 2 = 5" is both false and not true.

"2 + 2 > red" is neither true nor false because it is nonsense. If it were false, its negation "2 + 2 ≤ red" would be true, which is not the case.

• See also : not even wrong Oct 20, 2016 at 11:11
• While your distinction between falsehood and non-sense is overall correct, the second example you gave is dubious, since one might use logical atomism to show that is actually false: you can break down "2 + 2 > red" into (a) ^ (b) ^ (c) into (a) "2+2 is a number", (b) "red is a number" and (c) "The number from (a) > the number (b)". Since (b) is false, (a)^(b)^(c) is false as well. Oct 20, 2016 at 17:47
• @John Am: under classical truth-conditional logic "false" is definitely not synonymous with "nonsense". The former is a truth-value; the latter is not. there is no concept of "nonsense" in t-c logic; every proposition is either true or false. a nonsensical sentence is not a proposition, by definition.
– user20153
Oct 20, 2016 at 20:07
• I feel like you have to define "2+2>red" in a meaningful manner before you can legitimately talk about "negating" the statement. Though I certainly agree both the statement and its "negation" are undefined (under normal math) and therefore "not true" while also being "not false". Oct 21, 2016 at 3:57
• You don't even have to go so far as to get into nonsense. "How are you doing?" is neither true nor false for the same reason.
– hvd
Oct 22, 2016 at 10:23

In the classical logic something is neither true nor false if it is grammatically malformed to have a truth value, so 2+5 or "x is blue" are not "true", but not "false" either, they are not truth-apt. The classical assumption was that all truth-apt expressions can be distinguished by syntax alone, i.e. there is a clear way to tell from how they are formed whether it is truth-apt or not, without inquiring into what they mean. However, it is easy to come up with expressions that are grammatically well-formed but problematic semantically, sometimes crudely called gibberish, e.g. category errors like "electrons are blue". Those are also neither true nor false, at least intuitively. Wittgenstein even suggested that in natural languages there is no clear distinction between syntax and semantics, and there is no way to clearly prescribe what is well-formed, all rules are "grammar".

There are non-gibberish expressions that have problematic relation to the truth for other reasons, e.g. "such and such will win the election tomorrow". Is it already true (or false) today? Aristotle and modern intuitionists say "no". What about undecidable mathematical statements, like the continuum hypothesis? Same idea. There is also another dimension to the difference between true and false. The classical logic assumes for simplicity that that those are the only truth values that truth-apt expressions might take, this is called bivalence, often confused with the law of excluded middle. Multivalued logics remove this assumption. In particular, popular in applications fuzzy logic allows certain claims (usually "vague" ones) take any truth value between 0 and 1, with 0 being false and 1 true. So something like "15 degrees centigrade is cold" will be neither true nor false but have the truth value of say 0.6.

All of these phenomena led to the idea of logics with "truth value gaps", where we either interpret some expressions as having no truth value at all, or one different from "true" and "false". Sometimes we are forced to do this by the classical logic itself, e.g. the Liar sentence "I am false" leads to a contradiction if we assume that it has one of the two classical truth values. There is a whole field of semantic paradoxes like tha Liar, to resolve which Kripke specifically developed a whole semantic theory with truth value gaps. Paradoxes of vagueness, like the paradox of the heap (one grain is not a heap, adding a single grain won't make not a heap a heap, therefore no amount of grains makes a heap) can also be resolved using truth value gaps.

In classical logic these are the same by definition.

But in very tentative logics like Constructivism or Intuitionism, things are only said to be true or false if they meet quite stringent conditions. People using criteria like this require a truth to be proved in a given way, or captured by a certain kind of generalization, and a falsehood to proceed from a clear counterexample that meets the standard for truth. (The idea is that truth is ultimately negotiable, as our intuition improves, or that we should avoid claiming truths we cannot back up with computations.) That means that just not being false is not enough to make them true. There is a vast middle ground of things that remain inaccessible to truth or falsehood.

• I don't know what you mean by "very tentative", many serious mathematicians have studied both constructivism and intuitionistic logic.
– user23788
Oct 20, 2016 at 13:43
• I don't think this is intended to be pejorative so much to say that the logic systems by their nature accept a reduced set of deductive axioms, and thus are more "tentative" about accepting conclusions Oct 20, 2016 at 14:26
• truth-conditional (i.e. "classical") logic uses axioms. constructive/intuitionistic logic does not use axioms. there is nothing "tentative" about either.
– user20153
Oct 20, 2016 at 19:47
• I think you meant to say that constructive/intuitionistic logic also uses axioms. Well, that they're both formal deduction systems based on well-defined axioms and rules of inference. Oct 21, 2016 at 2:54
• 'Tentative' in the sense of withholding judgement. Intuitionism does not aggressively assign truth values when they are not needed.
– user9166
Jan 4, 2017 at 19:55

There is some ambiguity in what a person means precisely by the phrases.

For example, sometimes people use "P is true" (respectively "P is false") to mean that P can actually be proven (resp. disproven) in whatever logical system you're using.

With such a meaning, if P were an undecidable statement — one that can be neither proven nor disproven — then one would assert "P is not true" but not assert "P is false".

Similarly, if we assign truth values to propositions in a multi-valued logic, natural language doesn't do a good job distinguish between

• We did not assign the value "true" to the proposition P
• We assigned the value "true" to the proposition "not P"
• We assigned the value "not true" (i.e. "false") to the proposition P

so again it's somewhat ambiguous exactly what a person means if they say "P is not true" or "P is false"

In some cases not true could be either false or nil, but mostly not true just means false.

Truth is a condition of statements (utterances, propositions, sentences, and such - see chapter 9 of John R. Searle's "The Construction of Social Reality"). This condition is satisfied when utterance matches (fits, corresponds to...) what is (the case, the world, states of affairs, et cetera. The adjective "true" describes the satisfaction of this condition.

"Not true" and the synonymous adjective "false" describe a state of affairs where this condition of utterance is unsatisfied (met, obtained...) "Not true" is also used in a sense which "false" is not commonly accepted to indicate a nil (or null) status regarding truth value. For example, the statement is neither true or false that "Ruebens is a better painter than Pollock" - simply a matter of agreement, or what is commonly described as "true to you" (or me, or us, or true to them). Such are matter of sentiment, opinion, poetic use of language and such. Note that the objects of this sentence (Ruenbens, Pollock, their paintings, and the utterers opinion of them) are non-fictional. I mention this to distinguish a sentence with only perspectival (or situational) truth value (not empirical or axiomatic truth value) from sentences which has no rationally assessable truth value, such as "colorless green ideas sleep furiously" or "god did it" or "Yay!" or "go away".

This is to say that such sentences are not therefore falsehood, only that evaluating the truth condition is inadequate to the occaision of their utterance. Consider, even a sentence which is rationally assessable as true or false may in fact be uttered without concern for such an evaluation. For example, if I am in my shower while practicing French and reciting the sentence, "il pleut" it is not a comment upon microclimate conditions. If I enter a French cafe completely soaked from a sudden Summer rainstorm to which the patrons were unaware and upon receiving odd looks I motion upwards and utter, "il pleut!" then it is an explanation of my circumstance including a comment upon the weather which can be assessed a truth value (all the patrons need do is look outside and see that either it is still raining or that everything which was dry when they entered the cafe is now wet to empiricaly verify the state of affairs and render an accurate truth value to my utterance).

In a specialized discourse, "truthy" and "falsey" have broader use of "not true" but only in the technical arena and not for empirical statements of the case.