1884 -] Canine Intelligence, 11 
objects which are round ; we can form (and formulate) 
abstract ideas of numerical relations apart from the objedts 
which enter into these relations, and we can frame and 
discuss abstract ideas of figure apart from the objedts which 
exhibit this spatial relation. These abstradt ideas may, 
moreover, like our ideas of objects (groups of qualities in 
natural combination), be either particular or general. The 
abstradt idea of roundness, for example, is and must always 
be particular, for there are not different kinds of roundness 
which may give rise to a general idea of roundness. The 
abstradt idea of a triangle, on the other hand, maybe either 
particular, as in the case of an isosceles triangle, or general ; 
the general abstradt idea being a conception of that general 
relation to each other of three straight lines, without which 
no triangular figure can exist. 
Note now the difference between the general ideas of 
sweet, bitter, &c., to which Mr. Romanes alludes, and the 
abstradt ideas of sweetness, bitterness, and so forth. The 
ideas of sweet, bitter, &c., formed by a dog are not isolated ; 
they form part and parcel of the conception of the objedt 
under examination. They are general because no special 
form of sweetness, bitterness, &c., can be called up in asso- 
ciation with the other qualities of a hitherto unexplored 
objedt. But they are not “ apart from the objedt examined.” 
There is no separation in thought of the qualities of the 
objedt in question. 
This separation, which I hold to be essential to an abstradt 
idea properly so called, is, to quote the words of John Locke, 
“ an excellency which the faculties of brutes do by no means 
attain to.” It is only the human mind, and that mind in a 
somewhat advanced state, that can analyse an objedt (or 
group of qualities in natural combination) into its component 
parts, and then consider certain of those qualities to the 
exclusion of the others. 
Man in fadt does two things which no dog can do. By 
analysis he breaks up his conception of an objedt into sepa- 
rate components, and studies those components by them- 
selves ; and by synthesis he builds up a more comprehensive 
conception, adding fresh components which he has discovered 
by study and refledtion. It is by the symbolic use of words 
that he is enabled to carry on his analysis ; and, on the 
other hand, as I have said elsewhere, “ it is the word that 
groups around itself not only the little cluster of associated 
ideas which make up the ordinary unrefledting conception 
of the objedt it symbolises, but also all those further ideas 
which are the result of scientific study. The word is the 
