184 The Loom of Language
Through such culture-contacts words have wandeicd from one
language to another of a totally diilercnt ongin The modern word
bicycle pedals over hnguistic frontiers as the machine used to pedal over
national boundaries before passports were obhgatoiy The word-
material of all, or nearly all, languages is moie 01 less mongrel Imn in
the more exclusive members of the Teutonic gioup the number of
intruders is many limes larger than the number oi woids which the
linguist thinks he can trace back to the hypothetical common idiom
called puirutivc Teutonic When dealing with words for numbers, or
weights and measures, we have always to reckon with the possibility of
cultural, and thereibie wotd, diffusion II voeabuLuy is the only clue
available, we have to yve due consideration to geogiaphical situation.
If two languages which shaie a considerable poiuon of conseivauvc root-
words are not geographically contiguous, it is highly piobablc that they
are related
Word-similarity is a good clue* A second is agreement with respect
to grammatical be/iM tow Hench, Spanish and Italian, which we may
use as our control gj oup, have a host of common grammatical features
such as
(i) A tuturc tense (sec pp J06 and 339; which is a combination oi
the miimtive and the auxihuty to have (I-i aimct-ai^ aitncr~a<>3
Ital amar-o, amat-ai\ Spaa, amui-c^ unuiY-a\)
(nj The clciuutc article (Fi, masc, le> icm /a, Span cl 01 la> ItaL il oi
la) 3 and pronouns oJ the third person (Fr il or dlc} Span d or
elicit ItaL cgh or clUf) all derived horn the Latin demonstrative
ilk* ilia
(111) A twolold gcndex system m which the nuseulme noun generally
takes the place of the Latiu neuter (li'r /r wn9 the wme,
Span &l vino*} Ital it vino^ Latin mnutti)
Grammatical peculiarities, like woids, may be moie or less conserva-
tive. In the widest sense ol the term, grammar includes the study of
idiom and sentence construction, or yywtax, m contradistinction to
accidmcCy which deals with the modification of individual words by
flexion or root~vowel changes The syntax of a language is much less
conservative than its accidence. When we meet with resemblances of
the latter type, it would be far-fetched to attribute them to chance or
to borrowing* All the evidence available tends to show that, while words
and idioms Muse freely, peculiarities of acadmcc do not Now and
then a language may borrow a preta or a sulfix> together with a foreign
word, and subsequently tack one or the other on to indigenous words,