20 BEACH-LA-MAR. 
which has arisen to prevent the comprehension of isolating language 
has been the fetish regard in which students have held eight or nine 
parts of their own speech as in some sort deodand. 
This is not the place, nor in this treatise is there room, for a com- 
plete discussion of the syntactical problems of isolating speech. But 
our study of the jargon calls for at least a summary statement of the 
rules of Melanesian grammar in accordance with which it is used. 
All vocables fall into one of three classes. This is a present and 
operative condition; it is not impossible, in many cases it is easy, to 
study out the method of differentiation by which these classes have 
grown from a protoplasm of sounds modulated by a consciously 
exerted intelligence. The three parts of Melanesian speech, the 
designation familiar to our systems of grammar being conveniently 
employed in an indicative sense and not as definition, are the attrib- 
utive, the demonstrative, and the paradeictic. 
The list of the vocables which fall within the third class is brief. 
These words are such as indicate, but do not necessarily define, 
relation as existing between two objects of cognition. ‘The intellec- 
tual plane of the men whose thought is communicated by these 
languages is yet far too low to give to these relation words positive 
and distinctive value; their effective end in speech is no more than 
to indicate that at a certain point there is a relation of some sort. 
In this category we may discover the segmentation of the germ 
which seems to promise growth into something corresponding to the 
conjunctions and prepositions of the more highly organized systems 
of speech. 
There are very few of these paradeictic words—quite enough to 
serve the islander’s needs in distinguishing the several sorts of rela- 
tion which seem to him valuable to communicate. Hence it comes 
about that each of these few words must do recruit duty for a large 
number of the relations to comprehension of which our keener intel- 
ligence and recognition of a more deeply interlaced plexus of associ- 
ation have brought us. We note that these words are among the 
most elemental in the several languages; they are the simple vowels, 
or at most they have undergone the most primary modulation by the 
prefixing of a coefficient consonant. To exhibit both the nature and 
employment of the paradeictic we may profitably consider a colloca- 
tion of vocables drawn from one of the isolating languages of the 
Pacific, and I employ the Samoan of Nuclear Polynesia for the 
reason that it has been so much more extensively studied than any 
Melanesian speech that more detailed examination of the point is 
comparatively easy. 
In the composite sense group fala 1 manu we see two attributives 
between which is interjected the paradeictic i. If we were to render 
this phrase into Latin we should have istoria animalium; we should 
