GRAMMAR OF ISOLATING SPEECH. 21 
find the dependence of the succeeding attributive word upon the 
former paradigmatically incorporated within itself; we should parse 
animalium as in the genitive objective. Therein we confess (at least 
an examination of the logical process underlying the genitive use as 
object shows that without particular note of the fact we do acknowl- 
edge) that historia, though noun in its assignment to the classic parts 
of speech, yet retains so much of its verb power as to govern an 
object. In any English rendering we must employ a preposition; we 
can make the Samoan sense appear only through some such expres- 
sion as “‘story of animals” or “‘story about animals.’’ When we 
examine the Samoan dictionary—and we find the same thing in all 
Polynesia and in all Melanesia—it is observed that paradeictic 2 
is variously rendered as 72n, at, to, with, about, of, for, by, or eludes 
specific rendition entirely. This heaping up of significations should 
show on the first inspection that paradeictic 7 is not any one of these 
English prepositions, it is not all of them; it is still so elemental a 
part of primordial speech that it is far from becoming preposition at 
all. ‘The utmost that the Samoan conveys to his alert hearer, the 
utmost that he regards it essential to convey for purposes of thought 
communication, is the following, and here I must anticipate the 
explanation of the attributive which will be reached in due and 
orderly course. He says: “‘there-is-a-telling—there-is-a-relation— 
living-things.”’ Thus paradeictic 7 is not here a preposition; it is 
but a warning sign that in the former attributive is a verbal value 
and that it has transitive force upon the succeeding attributive. It 
is, therefore, but a sign suggesting a certain group of relations. 
Before we pass along and leave this particular paradeictic we 
may well note another of its uses, for it is critical in our classification 
of these languages as isolating and not agglutinative. To warrant 
a language entry into the agglutinative class it is not sufficient to put 
two elements into compaction. Clearing away the iron rule of the 
printer, such compaction is a matter of pronunciation. We acknowl- 
edge no such virtue in the nut-quad of the printer’s case as that 
by its presence or by its absence it shall make a word compound on 
the one hand or on the other composite. So, as between isolating 
languages freely compacting and agglutinative languages freely com- 
posite, there must be a difference underlying the distinction. This is 
it. In agglutination we encounter a modification and more or less 
of atrophy of the subordinate members of the composite. 
In the foregoing exhibition phrase from the Samoan we have 
already seen tala the attributive word and 7 the paradeictic. In both 
of the existing vocabularies of this language, and for motive of con- 
venience I shall retain it in my dictionary, we encounter as a second- 
ary attributive talaz, erroneously designated verb but fairly employed 
in no senses other than such as are comparable with our under- 
