6

I was thinking about if there is any difference, depending of the context, by saying "for each" and "for all" and I think I may found one context for this.

Considering the definition of Cauchy sequences:

It states that for every $\epsilon>0 \exists $ an $N$ st.. blablabla... Of course there's no unique $N$ which works for all $\epsilon>0$, because if there was it should be the maximum $N$ of all $\epsilon>0$ we considered. But then we could choose an smaller value of $\epsilon$ (let's say $\epsilon_1$), therefore $N(\epsilon_1)$ would be greater than the others, so there's no such thing as an unique $N$ that works for all $\epsilon>0$.

So, "for all" in this context means the same as "for each", because we cant have an $N$ which works for all $\epsilon>0$ at the same time, so we have for each $\epsilon>0$ an N st blbabla...

But suppose there exists an unique $N$ which works for all $\epsilon>0$. So, in this assumption no matter what $\epsilon$ you choose, the distance beetwen the elements of the sequence whose indices are greater or equal than $N$ would be always less or equal than $\epsilon$. So for all $\epsilon>0$ there exists an unique $N$ st blablabla...

Now, if I say for each $\epsilon>0$ there exists an unique $N$ st blablabla... It just states that exists an unique $N$ that works for each $\epsilon>0$ but it doesnt ensures that $N$ works for all $\epsilon>0$, so it has a different meaning by saying "for each" and "for all" in this context. What do you think about that??

Of course my assumption is false, but there could be an similar context which is true, and by using the same reasoning, "for each" and "for all" would have different meaning.

user312479
  • 473
  • 1
    In "There exists an $N$ such that for all $\epsilon > 0$", $N$ is stated without knowing what $\epsilon$ is. Conversely, "For all $\epsilon > 0$, there exists an $N$" states $N$ after $\epsilon$. – Henricus V. Mar 09 '16 at 01:36
  • 1
    Well, there is sometimes a difference: sometimes one of the words fits in an English sentence where the other word would seem awkward or unnatural. In such a case you will use the word that is better suited to regular English usage. But in that case it is not true that the other word can be used to say something different; you simply should not use the other word for any reason in such circumstances. That's clearly not the kind of distinction you're trying to make, so I didn't propose it as an answer. – David K Mar 09 '16 at 02:20
  • @Henry in the first case N would not be function of $\epsilon$ (of course this is not true)... but assuming this is true, then "for each" and "for all" would have different meaning (if my reasoning was right). Actually, this could be a ''proof by contradicition'' that "for each" and "for all" it's the same thing. lol – user312479 Mar 09 '16 at 02:27
  • 3
    I'm voting to close this question as off-topic because it has degenerated into an English lesson. – Rob Arthan Mar 09 '16 at 02:56

4 Answers4

6

So, if I understand you, you are suggesting that the keywords "each" and "every" should map to the order in which we write mixed universal and existential quantifiers, regardless of the spoken order.

  • "There's some $y$, that makes it true, for every $x$" should be interpreted as: $~\exists y~\forall x~\operatorname{It}(x,y)$ .
  • "There's some $y$, that makes it true, for each $x$" should be interpreted as: $~~\forall x~\exists y~\operatorname{It}(x,y)$ .

Well, it would be useful if naïve speech was that consistent in the use of "for each" and "for every", and if everybody always used "for all" for one rather than the other.   Unfortunately that is counterfactual; it can't be relied on.   There are enough exceptions to always require double checks on the speaker's intent.

It would be far less confusing and error prone to encourage students to learn to use the same order in both speech and symbols.

  • "There's some $y$, that makes it true, for all $x$" should be interpreted as: $\boxed ?$ .
Graham Kemp
  • 129,094
  • 1
    (ignoring the OP's obsessions about "each" v. "all") it is devastatingly confusing to do what you did and write quantifiers before and after the statement they quantify. – Rob Arthan Mar 09 '16 at 03:01
5

There is one context in which "all" and "each" cannot be interchanged, and that is when describing an algorithm of some sort.

Put half of each of the fruits into the red box and the other half into the blue box.

Put half of all the fruits into the red box and the other half into the blue box.

Following the first results in each fruit being split up into two halves that go into different boxes.

Following the second results in the fruits being divided into two groups that go into different boxes.

Similarly consider:

There is an upper bound for each function in the sequence.

There is an upper bound for all functions in the sequence.

The first states that the each function in the sequences has an upper bound, while the second is usually taken to state that there is a common upper bound for all the functions in the sequence. (Here it is irrelevant what exactly "upper bound" means.)

Since Rob Arthan considers the second sentence in the above example as ambiguous, here is a much clearer pair:

There is a unique cardinality for each set.

There is a unique cardinality for all sets.

And for people who already know too much about cardinality and hence their interpretation is biased, here is an even clearer example:

There is one teacher for each student.

There is one teacher for all students.

The moral of this is to use "each" whenever referring to each one individually and to use "all" whenever referring to all of them as a whole collection (IF one insists on writing in natural language with quantifiers on both sides of the statement). In many cases they are interpreted to mean the same thing, but not always, as I showed.

If there are two or more quantifiers, I personally recommend putting all the quantifiers in front. Sometimes that sounds a bit unnatural in English, in which case one should use commas and add words like "common" or "same" and choose carefully between "each" or "all" to ensure that misinterpretation is impossible.

Community
  • 1
user21820
  • 57,693
  • 9
  • 98
  • 256
  • What is your evidence for the claim that "there is an upper bound for all functions in the sequence" is usually take to mean there is a common upper bound? To me, it's just sloppy writing with no clear meaning. – Rob Arthan Mar 09 '16 at 23:34
  • @RobArthan: I'm actually very strict when it comes to precise writing, but you cannot deny that many mathematical writings are as sloppy or worse than the pair of sentences in my second example. Nevertheless, I added a third example that I think you'll agree is relatively precise (by most mathematicians' standards, not mine) and yet shows a difference between "each" and "all" despite being intended for quantification. – user21820 Mar 10 '16 at 01:22
  • I think this points out the challenge with potentially unclear writing. To me, "there is an upper bound for all functions in the sequence" is much more understandable than "there is a unique cardinality for all sets". The former seems to say that the sequence is uniformly bounded; the latter is painful to read, but the only meaning I can give it is that each set has a cardinality. – Carl Mummert Mar 10 '16 at 01:44
  • @CarlMummert: Yes indeed I believe many people would read the "upper bound" example to involve uniform boundedness, which is why I had it as my original example. I think your pain with the "unique cardinality" example is due to your prior knowledge of cardinalities. Imagine that "cardinality" is some kind of unknown label, and it is not known what kind of labels a set might have, and then it would become natural to interpret the second sentence as saying that every set has the same unique label. I better edit to clarify my stand on precision... – user21820 Mar 10 '16 at 01:53
  • @CarlMummert: I have thought of and added an unambiguous example (in standard English) which is very similar to the cardinality example. – user21820 Mar 10 '16 at 02:42
  • @RobArthan: I've added a fourth example that is unambiguous in English and verifies my claim that the choice of "all" or "each" may result in two unambiguous sentences that actually has quantifiers switched. Also, I added the last paragraph to clarify my opinion on mathematical writing, about which I'm in agreement with you that we should be precise enough and leave no room for ambiguity. Do you agree now? – user21820 Mar 10 '16 at 02:47
  • These give an interesting example. I think the real key is to these examples is that the actual predicate (i.e."teacher") is being placed in the middle of the quantification (before for "for all students"), and that the actual objects being quantified are not given variables, which obscures the predicate even more. If we write the English more like a formal quantification, the meaning of "each" is indeed the same as the meaning of "all": "there is a T such that for each S, T is a teacher of S". But, when written as in the answer, the meaning does appear to be different. – Carl Mummert Mar 10 '16 at 11:19
  • 1
    @CarlMummert: Actually I dispute your analysis. It would be the same even with the quantifying words in front; compare "All students have one teacher." and "Each student has one teacher.". My explanation is simply that "all X" is treated as a group by default and only split into individuals when context or prior knowledge demands it (that is also why the cardinality example did not satisfy you), whereas "each X" of course refers to each individual. But yes, I usually write mathematical statements in a proof in rather mechanical (non-idiomatic) English. =) – user21820 Mar 10 '16 at 11:50
  • I would read "all students have one teacher" as a synonym for "each student has one teacher". But I also think it is interesting that these only have one quantifier, over students, they do not quantify over the teacher at all. So they still do not explicitly show all the quantification. Once we write out the quantifier over the teacher, it will be more obvious if it is before or after the quantifier for the student. – Carl Mummert Mar 10 '16 at 11:53
  • @CarlMummert: Interesting; I did not think anyone would read "all students have one teacher" as the same with "each". I suspect that dialectical differences is playing a large role here. I recall observing before that American English tends to use "all" even when it would be considered incorrect by British speakers, but I didn't know it extended to my example. However, I'm sure you'll agree that "All the students have only one teacher" is not the same as "Each of the students has only one teacher". Heheheh. – user21820 Mar 10 '16 at 12:36
  • @CarlMummert: I'm not sure we can say that there is no quantification. Consider "All the students in this school are taught by a few teachers." and "Each of the students in this school is taught by a few teachers.". Does "a few" not quantify "teachers" in your opinion? Linguistically it is counted as a quantifier. And how about "Only a few teachers teach all the students." (a bit ambiguous) and "Only a few teachers teach each student."? =P – user21820 Mar 10 '16 at 12:43
  • You aren't parsing your latest examples right. "There is one teacher for all students" is not using "for all" as predicate quantifier. "one teacher for all students' is noun-phrase acting as description of a teacher. – Rob Arthan Mar 10 '16 at 21:08
  • @RobArthan: You're mistaken. "for all X", whether or not it is front, functions as an adverbial phrase, because "for" is a preposition. It does not matter if it used to modify a noun phrase; "apple on the table" is the same as "apple that is on the table". Same for "teacher for all students". Besides, you cannot deny that the only way to translate it is to use the quantifier, regardless of your linguistic interpretation of the English construction. And even if you disagree with my last example in my answer, you can't reject the examples in my comment above yours. – user21820 Mar 11 '16 at 03:16
  • @user21820: Consider the noun-phrase "a man for all seasons". "There exists a man for all seasons" means something quite different from "For all seasons, there exists a man". You may find Russell's essay on denoting interesting. – Rob Arthan Mar 12 '16 at 23:32
  • @RobArthan: I know that those are different. But it does not nullify my point at all. "a man for all seasons" expands to "a man that is for all seasons", where "is" is further interpreted according to context, such as "succeeds". So of course "there exists a man for all seasons" means "$\exists m \in man\ ( \forall s \in season\ ( succeeds(m,s) ) )$", unlike "for all seasons, there is a man" which means "$\forall s \in season\ ( \exists m \in man\ ( livesduring(m,s) ) )$". Your example supports my point, because the English word order causes quantifier switch, interpretation of "is" aside. – user21820 Mar 13 '16 at 02:33
  • @RobArthan: Also, you still have not responded to my examples in my third and fourth comment above this one, which were before your objection, which is unimportant if you cannot reject these examples. – user21820 Mar 13 '16 at 02:36
1

"For each" and "for all" are synonymous. However, the order in which you interleave "for each" (or "for all") and "for some" makes a difference and you can make life very hard for yourself if you make the mistake of putting one at the beginning and one at the end:

  1. for each $\epsilon > 0$, for some $N> 0$, $1/N < \epsilon$
  2. for some $N > 0$, for each $\epsilon > 0$, $1/N < \epsilon$
  3. for each $\epsilon > 0$, $1/N < \epsilon$, for some $N > 0$

1 is true, 2 is false and 3 is ambiguous. (I think you were thinking of a sentence along the lines of 3 when you began to think that "for each" and "for all" were not synonymous.)

Rob Arthan
  • 48,577
  • I know that the right thing to do it's 1). But I assumed the false stating that exists one N which works for every $\epsilon>0$ on the context of Cauchy sequences. So in this case, I can reverse order and 2) (changing ''for each" that you wrote to "for all") would be true. However, this case do not exists (in the context of Cauchy sequences at least), but if there's some context that my assumptions is true, then would exist some $N>0$ st for all $\epsilon>0$, $1/N<\epsilon>0$ and this would not be right if I interchange "for all" to "for each", thus in this context they have different meaning. – user312479 Mar 09 '16 at 02:20
  • But, whathever, this is not important and it's quite confusing. – user312479 Mar 09 '16 at 02:21
  • No: you can change "each" to "all" throughout my examples and it makes no difference to the meaning. – Rob Arthan Mar 09 '16 at 03:02
  • @RobArthan: Your claim that "for each" and "for all" are synonymous is not really right; see my answer. Yes this is a problem arising solely from idiosyncracies of English, but it's directly related to mathematical writing and so I think it deserves a place here. – user21820 Mar 09 '16 at 11:53
  • @User21820: your answer points out a twist where "for each" and "for all" when not used as logical quantifiers are not synonymous. That is not relevant to the question as asked. – Rob Arthan Mar 09 '16 at 23:48
  • @RobArthan: I presume you're referring to the first part of my answer in your above comment. I disagree that it's irrelevant; description of algorithms is very much relevant to mathematics, and "each" and "all" do correspond to the universal quantifier, just that they have different binding strength. Construction of a function literally is construction of something that satisfies a universally quantified statement about every object in its domain. – user21820 Mar 10 '16 at 01:10
0

The key point in mathematical English is that the order of the quantifiers in English should match the intended formal order. So, if we write out the quantifiers very explicitly:

  • There is an $x$ such that, for all $y$, $P(x,Y)$ is $$(\exists x)(\forall y)P(x,y)$$

  • There is an $x$ such that, for each $y$, $P(x,y)$ is also $$(\exists x)(\forall y)P(x,y)$$

  • For all $y$ there is an $x$ such that $P(x,y)$ is $$(\forall y)(\exists x)P(x,y)$$

  • For each $y$ there is an $x$ such that $P(x,y)$ is also $$(\forall y)(\exists x)P(x,y)$$

As you can see, "all" and "each" are synonymous in that sense. The order is all that matters, if things are written out in enough detail (the "such that" phrasing).

In more general English phrasing, without the "such that", it can be very hard to tell exactly what is meant, as the answer by user21820 shows. In some cases, you need to have enough context to already know what is intended (or be able to figure it out).

There is one more issue: many authors will write the innermost quantifier after the body of the formula, like this:

  • There is an $x$ such that $P(x,y)$ for all $y$ is $$(\exists x)(\forall y)P(x,y)$$ for many authors.

This convention is not universal, and it can be confusing if you don't already know what is intended, but it is common enough to point out (especially in analysis with the definitions of limits and continuity).

Carl Mummert
  • 81,604