130 The Loom of Language
adjective when fonnally distinguishable. This is also true of all lan-
guages included in the Indo-European group. Similar remarks apply
with equal force to the pronoun. When we recognize as such a word
which lacks the characteristic terminals of an adjective, a noun, or a verb
in a flexional language like Latin, we depend laigely on the context For
instance, the English particles a or the are signals that the next word is
not a verb or a pronoun, and the presence of a pronoun usually labels
the next word of a plain statement as a verb A pronoun usually stands
for some name-word previously mentioned , but in certain contexts
personal pronouns may stand for anything which has gone before, and
it has no specific reference to anything at all, when used in what gram-
marians call impersonal constructions such as it seems. Neither the
pronoun nor the verb, which we recognize as such by the flexional ~s m
the same context as the third person ity here fits into any tidy definition
based on the fraction of words m a sentence, i e what they mean Few
of us now postulate a force not of ourselves which makes for raininess,
when we say it rains.
To some extent we select one of several word-forms with the same
general meaning in accordance with the process of analogical extension
which plays such a large part (p. 204) m the growth of speech. In
literate communities grammaiians also take a hand m shaping die
conventions - of language by prescribing certain patterns of expression
based on picccdcnts established by authors of repute, or on paradigms
from the practice of dead languages which have more ostentation-value
than vernacular utterance. The most time-honoured model of ttm type
is called the subject-predicate relation (see p, 117),
Till recently grammar books used to say that every sentence has to
have at least two components, a verb and its subject, which must either
contain a noun or be a pronoun Accordingly, it is incorrect to write
rainy day, what? The only intelligible definition wliich usually tells us
what grammanans would call the subject of a Latin or Greek sentence
is that it answers the questions formed by putting toko or what in front
of the vcib, and this does not get us far when we replace the preceding
expression by the "sentence", r* it not a rainy day? Who of what rains,
in this context, is less a matter of grammar than of theological opinion,
Buddhists and Christians., atheists and agnostics, would not agtte
about the coriect answer, and a Scots schoolmistress of any persuasion
would find it difficult to convince a Chinaman that the meaning of the
ensuing remarks would be more explicit if we put it i$ in front of the
first, and there is in front of the second: