Secure Epistemic Foundations, Language, and Reality

This from Grigory Aleksin:

I have been doing some reading and thinking, and there are a few things that I cannot quite get my head around. I was wondering whether you could help me, or point me in the direction of some work on the issue. My somewhat naive task has been to try and find the most foundational and basic pieces of knowledge that are required by any worldview. 

It seems to me there are at least two things that are in some sense foundational:

(1) Something exists

(2) There are correct and incorrect inferences

(1) seems to follow from what is meant by a 'thing' and what is meant by 'exists'. However this is only the case, if there are correct and incorrect inferences. Therefore, (2) is in some sense prior to (1). Hopefully that makes sense.

BV: It does indeed make sense. But I would approach the quest for secure foundations more radically.  How do I know (with objective certainty) that something exists? I know this because I know that I exist.   'Something exists' follows immediately from 'I exist.'  To say that one proposition follows from another is to say that the inference from the other to the one is correct.  The correctness of the inference preserves not only the truth of the premise but also its objective certainty.  I agree that your (2) is in some sense prior to (1); it is a presupposition of the inferential move from 

(0) I exist

to

(1) Something exists.

My problem arises when I consider that both (1) and (2) are not actually part of reality: both are sentences or linguistic expressions.

BV: Here you have to be careful. Surely a sentence token is a part of reality, even if you restrict reality to the spatio-temporal. The truth that something exists is not the same as its linguistic expression via the visible string, 'Something exists.' That same truth (true proposition, true thought) can also be expressed by a tokening of the German sentence 'Etwas existiert' and in numerous other ways. This suffices to show that the proposition expressed is not the same as the material vehicle of its expression. And already in Plato there is the insight that, while one can see or hear a sentence token, the eyes and the ears are not the organs whereby one grasps the thought expressed by marks on paper or sounds in the air.

So we need to make some distinctions: sentence type, sentence token, proposition/thought (what Frege calls der Gedanke). And this is just for starters.

And should we restrict reality to the spatio-temporal-causal? Are not ideal/abstract objects also real?  The sign '7' is not the same as the number 7. A numeral is not a number. I can see the numeral, but not the number. I can see seven cats, but not the (mathematical) set having precisely those cats as members. I can see the inscription '7 is prime' but not the proposition expressed on an occasion of use by a person who produces a token of that linguistic type.  The ideal/abstract objects just mentioned arguably belong to reality just as much as cats and rocks. 

Thus I have come to consider the role of language. The issue is that language is just a way of mapping reality, and as such is disconnected from it. This raises the question of what 'truth' is, since on one hand we know that there are objective truths, yet truths are only expressed [only by] using language. My question is, then: how can the analysis of language be used to answer philosophical questions? I know that linguistic analysis plays a central role in analytic philosophy, but I cannot help by having [but have] doubts or suspicions that something is wrong. As you see, I cannot fully express what it is that causes me such a headache, but it stems from a suspicion with respect to the use and limits of language, and thus philosophical inquiry. 

BV: We do distinguish between WORDS and WORLD, between language and reality. But this facile distinction, reflected upon, sires a number of puzzles.  My cat Max is black. So I write, 'Max is black.'  The proper name 'Max' maps onto Max. These are obviously distinct:  'Max' is monosyllabic, but no animal is monosyllabic.  So far, so good. But what about the predicate 'black'?  Does it have a referent in reality in the way that 'Max' has a referent in reality?  It is not obvious that it does.  And if it does, what is the nature of this referent?  If it doesn't, what work does the predicate do? And then there is the little word 'is,' the copula in the sentence.  Does it have a referent? Does it map onto something in reality the way 'Max' does? And what might that be?  The transcendental unity of apperception?  Being?  If you say 'nothing,' then what work does the copula do?

One can see from this how questionable is the claim "that language is just a way of mapping reality . . . ."  We don't want to say that for each discrete term there is a one-to-one mapping to an extralinguistic item.  That would be a mad-dog realism.  (What do 'and' and 'or' and 'not' refer to?) Nominalism is also problematic if you hold that only names refer extralinguistically.  And you have really gone off the deep end if you hold that all reference is intralinguistic.

Here is another ancient puzzle.  A sentence is not a list. 'Max is black' is not a mere list of its terms. There is such a list, but it cannot 'attract a truth-value.' That is a philosopher's way of saying that a list cannot be either true or false. But a sentence in the indicative mood is either true or false. Therefore, a sentence in the indicative mood is not a list.  Such a sentence has a peculiar unity that makes it apt to be either true or false. But how are we to understand that unity without igniting Bradley's regress?

And then there is the question of the truth-bearer or truth-vehicle. You write above as if sentences qua linguistic expressions are truth-bearers.  But that can't be right. How could physical marks on paper be either true or false? 

My question is, then: how can the analysis of language be used to answer philosophical questions?

It is not clear what you are asking. You say that there are objective truths. That's right. Your problem seems to be that you do not see how this comports with the fact that truths are expressed only by using language.  The source of your puzzlement may be your false assumption that sentence qua linguistic expressions are the primary vehicles of the truth-values.

Combox open.

Comments

[AleksinGrigory]

About my last question. Consider the ontological issue of holes. The sentence, "The piece of paper has a hole it in it" can be paraphrased in a number of different ways. For instance,

∃x(x is a piece of paper & x is singularly perforated)

∃x∃y(x is a piece of paper & y is a hole in x)

Notwithstanding those less that perfect translations, I do not understand how, or perhaps why, these manipulations ought to contribute to our philosophical understanding of holes? On the face of it, they undermine they introduce a sort of relativism in our use of language, and provide evidence to the claim that language does not capture reality very accurately. If that is that case, how then can we use language to give an accurate expression of metaphysical truths?

[David Gudeman]

Do analytical philosophers really think that analysis of language can answer philosophical problems? Problems of language, certainly, and maybe certain problems involving human cognition, but not ontological or epistemological or normative problems--as far as I know.

Analytical philosophers have used analysis of language to undermine philosophical positions. For example:

(1) In a sentence such as "In Germanic folklore Odin exists" it is a mistake to think that "Odin" refers to anything or that "exists" is meant to modify the thing that "Odin" refers to, so to the extent that Meinong's ontology is based on this mistake, it is unjustified.

Arguments of this sort are often made in service to competing metaphysical claims, but the argument itself doesn't lead to a metaphysical claim; it only undermines a metaphysical claim.

Are there actual metaphysical claims that are made based on language analysis? I can't think of any.

[AleksinGrigory]

This I think will be the last addition to what I wrote before. Hopefully, it will be interesting to others too.

Although I still do not know, what it is that I do not know, here is an attempt at a presenting the problem. Consider the following:

(0) There are correct and incorrect inferences

(1) I exist

(2) Something exists

As put above, we know can infer correctly (2) from (1). This assumes that there is a distinction between inferences that are correct and incorrect. Thus there are 'laws of logic'. However this is made more complicated by the fact that there is a distinction between true proposition and tru propositions that I know. It seems to me that the following is true:

Language Thesis: For every x that is a true proposition, x has to be expressed in language to be known

This seems true enough. When I say, 'Snow is white' or in Russian 'Cнег белый', I express the same meaning, but that 'meaning' is itself only understood using language. The meaning of the sentences 'Snow is white' & 'Cнег белый' is that "Snow is white". The meaning is only known through language. From this it would follow that for me "know" the proposition that 'I exist', I would have to presuppose knowledge of language. Equally, if I were to know (0), then I would only know it if it were expressed using language. However here comes the problem. Language surely makes some logical presuppositions, perhaps identity statements of sorts. However, we now have a circularity problem:
in order to know the logical presuppositions of language, expressed as propositions, I have to know language first. To put my line of thought more clearly:

(1) For every x that is a true proposition, x has to be expressed in language to be known (premise)

(2) 'I exist' and 'Something exists' are true propositions (premise)

(3) For the propositions 'I exist' and 'Something exists' to be known, the propositions have to be expressed in language (by 1 & 2)

(4) 'Something exists' is a true proposition because it is correctly inferred from 'I exist' which is also a true proposition (premise)

(5) If any proposition y is inferred from a proposition z, then logically prior to the inference there are true propositions that are 'laws of logic' or 'rules of inference' (premise)

(6) Hence, logically prior to the inference there are true propositions that are 'laws of logic' or 'rules of inference' (by 4 & 5)

(7) Hence, logically prior to the inference (from 'I exist' to 'Something exists'), and for the 'laws of logic' to be known, they a have to be expressible using language (by 1 & 6)

(8) There is a set of propositions or 'laws of logic' A, every member of which must be true, in order for language to meaningful (premise)

(9) In order to know that language is meaningful, every member of the set of propositions A must be known (by 8)

(10) Thus, in order to know that language is meaningful, every member of the set of propositions A must be expressible using language (by 9 &1)

However thinking about it (10) is absurd because in order to know that language is meaningful, I have to assume that is meaningful. Hopefully, I have not made too many errors in the above. I still have a lot of study ahead of me. Admittedly, your blog post on "The Discursive Framework, Logic, and Whether the Via Negativa is the Path to Nowhere" changed my attitude towards language and now I cannot stop thinking about it.

[Dmitri]

@Grigory: regarding your "...task... to try and find the most foundational and basic pieces of knowledge that are required by any worldview" and the question "how can the analysis of language be used to answer philosophical questions?" that you have repeated upon reading Bill's response. I suggest first to re-read the response as it is both brief and informative and does not presuppose technical philosophical knowledge.

If that does not work, I'd recommend Russell's "Problems of Philosophy" and (not or!) Nagel's "What Does It All Mean?" -- both good short introductions that could help you to come further on the path of achieving your goal.

And consider the following incomplete analogy: there are many pictures, depictions and descriptions of, say, Pisa Tower. As long as these different representations are of Pisa Tower, they are all true. Some are better, some are worse relative to your goals. Some are relevant to what you are looking to know about the tower. Some false representations, in the sense discussed here, could be relevant to your task even more than the true ones.

[AleksinGrigory]

Thanks everyone. There is still more thinking and studying to do for me, so I appreciate the comments and suggestions.

[BV]

You're welcome, Aleksin. Best wishes.

[Oz]

>>And you have really gone off the deep end if you hold that all reference is intralinguistic.

Did you have any theory in mind here?