Def Distribution/Linguistics/Lyons: each linguistic unit has a characteristic distribution, namely the set of contexts in which it can stand.
Distribution equivalent: two expressions can be in the same context. Correspondingly distribution-complementary or distribution-overlapping.
If two units are at least partially distribution equivalent, they cannot contrast with each other.
Distribution/Grammar/Lyons: We can take distribution as a starting point for a grammatical description: expressions have meaning when they are used in an appropriate context.
Distribution analysis/grammar/Lyons: a list would not be the most direct description of a text: in a sufficiently large language sample there will be a considerable overlap in the distribution of different words.
Distribution classes: Classes of words that can be used for each other in a sentence. Example "I drink beer": liquor, milk, water... and so forth
General/formal: e.g. we assume a material corpus of 17 "sentences": ab,ar,,pr,qab,dpb,aca,pca,pcp,qar,daca,dacp,dacqa,dacdp,acqp,acdp.
Letters: stand here for words.
Problem: we still have no distinction between "grammatically correct" and "meaningful" (useful).
In our example, a and p have certain contexts in common (namely -r,pc-, dac-), b and r (a-,qa) and d and q (dac-a,-aca, ac-p)
c: is unique in its distribution (a-a,p-c,p-p,qa-a,da-a,da-p), because no other word can be found in the same context as c.
X: we now merge a and p in the class CX and insert this class name everywhere where either a or p occur. Sentences that differ only in that o is where the other sentence has a are thus grouped into a class. ((s) "disjunctive"): Xb,Xr,(ar,pr), qXb,dXb,XcX, (aca, pca, pcp), qXr, qXcX,dXcX (daca,dacp), dXcqX,dXcdX,qXcdX,XcdX,XcqX,XcdX,XcdX.
Y: we set Y for b and r,
Z: for d and q.
Then we get
1. XY, (Xb,Xr)
2. ZXY (qXb, qXr, dXb)
4. ZXcX (qXcX, dXcX)
5. ZXcZX (dXcqX, dXcdX, qXcdX)
6. XcZX (XcqX, XcdX).
N.B.: with this we can capture the 17 sentences of our corpus through six structural formulas. (c is a one-membered class). They specify which consequences of word classes are acceptable. The consequences are linear. (see below).
Grammatically correct: are sentences in our example, that result from these structure rules. This is only achieved by the fact that the sentences that occur are regarded as links in a superset of 48 sentences. (The number 48 is obtained by applying the syntagmatic length formulas (see I 82 above) to each of the six sentence types and adding the results).
Generative/generative grammar/Lyons: the "grammar" in our example is generative in that it assigns a certain structural description to each sentence that appears in the "sample", for example pr is a sentence with the structure XY, pcda is a sentence with the structure XcZY, etc.
Grammar/Lyons: as it is understood here, it is nothing else than the description of the sentences of a language as combinations of words and word groups due to their affiliation to distribution classes. It is a kind of "algebra" in which the variables are the word classes and the constants or the values assumed by the variables in certain sentences are the individual words._____________Explanation of symbols: Roman numerals indicate the source, arabic numerals indicate the page number. The corresponding books are indicated on the right hand side. ((s)…): Comment by the sender of the contribution. The note [Author1]Vs[Author2] or [Author]Vs[term] is an addition from the Dictionary of Arguments. If a German edition is specified, the page numbers refer to this edition.
Semantics Cambridge, MA 1977
Introduction to Theoretical Lingustics, Cambridge/MA 1968
Einführung in die moderne Linguistik München 1995