VII (a) 15
Syntax/Quine: their rules are meaningful in contrast to their notation.
- - -
VI 69
Syntax/translation/indeterminacy/Quine: many of my readers have mistakenly assumed that uncertainty also extends to syntax. There was a subtle reason for this: in word and object
(1) (pp. 107, 129 136) it says:
VI 70
that also the specific apparatus of reification and object reference, which we make use of, is subject to indeterminacy. To this apparatus belong the pronomina, the "=", (equal sign) the plural endings and whatever performs the tasks of the logical quantifiers.
But it is wrong to assume that these mechanisms belonged to syntax!
>
Equal sign, >
Quantifiers, >
Pronouns, >
Indeterminacy.
VI 97
Spelling/Quine: resolves the syntax and lexicon of each content sentence and merges it with the interpreter's language. It then has no more complicated syntax than the addition sign.
1. Quine, W. V. (1960). Word and Object. MIT Press
- - -
VII (a) 15
Syntax/Quine/Goodman: their rules are meaningful as opposed to the notation itself.
- - -
XI 114
Language/Syntax/Lauener: Language cannot be regarded purely syntactically as the set of all correctly formed expressions, because an uninterpreted system is a mere formalism. ((s) This is not truthful).
XI 116
Lauener: it is a mistake to think that the language contributes the syntax but the theory contributes the empirical content. Therefore, one cannot say that an absolute theory can be formulated in different languages, or vice versa, that different (even contradictory) theories can be expressed in one language.
XI 136
Mathematics/QuineVsHilbert/Lauener: Mathematics is more than just syntax. Quine reluctantly professes Platonism.
- - -
XII 58
The problem of the inscrutability of the reference reaches much deeper than that of the indeterminacy of the translation: e.g. protosyntax.
>
Inscrutability.
Protosyntax/Uncertainty/Quine: the language here is a formalized system of proof theory of the first level, whose subject area consists only of expressions, i.e. of character strings of a certain alphabet.
Expressions: are types here, not tokens! (no occurrences).
Each expression is the set of all its occurrences. (Summarized due to similarity of inscriptions).
For example, the concatenation x^y is the set of all inscriptions that consist of two parts. These parts are tokens of x and y.
Problem: it can happen that x^y is the empty set ((s) the combination does not occur) although both x and y are not empty.
XII 59
The probability of this problem increases with increasing length of x and y!
N.B.: this violates a law of protosyntax that says:
x = z, if x^y = z^y.
Solution: then you will not understand the objects as sets of inscriptions.
But then you can still consider its atoms, the single characters as a set of inscriptions. Then there is no danger that the set is empty. ((s) Because the atoms have to be there, even if not every combination).
N.B.: instead of interpreting the strings as sets of inscriptions, they can be regarded as a (mathematical) sequence (of characters).
Character String/Expression: is then a finite set of pairs of a sign and a number.
Vs: this is very artificial and complicated.
Simpler: Goedel numbers themselves (the characters disappear).
Problem: Question: How clear is it here that we have just started to talk about numbers instead of expressions?
The only thing that is reasonably clear is that we want to fulfill laws with artificial models that are supposed to fulfill expressions in a non-explicit sense.
- - -
XIII 199
Syntax/Quine: "glamour" and "grammar" were originally one and the same word.
XIII 200
Later, the meaning also included magic.
Grammar: (in the narrower sense) said which chains of words or phonemes were coherent and which were not. Always related to a particular language.
Grammar: (wider sense): "The art of speaking" (in relation to the established use).
>
Grammar.
Syntax/Quine: for the narrower sense we do not really need the word "grammar", but "syntax". It is about which character strings belong to the language and which do not.
Problem: this is indefinite in two ways:
1. How the individuals are specified (formally, by components or phonemes) and
2. What qualifies them for the specification
XIII 201
Recognizability is too indeterminate (liberal).
Problem: ungrammatical forms are used by many people and are not incomprehensible. A language that excludes these forms would be the dialect of a very small elite.
Problem: merely possible utterances in imaginable but not actual situations that are not themselves linguistic in nature.
Solution:
Def ungrammatic/William Haas/Quine: a form that would not make sense in any imaginable fictitious situation.
Rules/Syntax/syntactic rules/Quine: are abstractions of the syntactic from long practice. They are the fulfillment of the first task (see above) to recognize which chains are grammatical.
XIII 202
Solution: this is mainly done by recursion, similar to family trees. It starts with words that are the simplest chains and then moves on to more complex constructions. It divides the growing repertoire into categories.
Parts of speech/Quine: there are eight: Nouns, pronouns, verb, adjective, adverb, preposition, conjunction, sentence.
Further subdivisions: transitive/intransitive, gender, etc. But this is hardly a beginning.
Nomina: even abstract ones like cognizance (of) and exception (to) are syntactically quite different, they stand with different prepositions.
Recursion/syntax/Quine: if we wanted to win the whole syntax by recursion, it would have to be so narrow that two chains would never be counted as belonging to the same speech part, unless they could be replaced in all contexts salva congruitate.
>
Recursion.
Def Replaceability salva congruitate/Geach/Quine: preserves grammaticality, never returns ungrammatical forms.
VsRecursion/Problem: if speech parts were so narrowly defined, e.g. Nomina, which stand with different prepositions, they would then have to be counted among different kinds of speech parts. And these prepositions e.g. of and to, should not fall into the same category either! Then there would be too many kinds of speech parts, perhaps hundreds. Of which some would also be singletons ((s) singletons = categories with only one element).
Solution: to give up recursion after having the roughest divisions.
