A. MACFARLANE—EXACT ANALYSIS AS THE BASIS OF LANGUAGE. 5 
ON EXACT ANALYSIS AS THE BASIS OF LANGUAGE. 
By A. Macfarlane, D. Sc., LL. D., Austin, Texas. 
Read February 6th, 1892. 
In recent years the invention of Volapuk has drawn the attention of 
thinkers to the problem of constructing an artificial language, and to 
the various solutions which have been attempted. In 1668 Bishop 
Wilkins published in the Philosophical Transactions of the Royal So¬ 
ciety of London a scheme founded on a philosophical basis. He first 
of all attempts to make a complete enumeration and description of all 
that is or can be known, and then makes this dictionarv of notions the 
basis of a corresponding dictionary of signs, both written and spoken. 
Such basis is necessarily a foundation of sand, for it supposes our 
knowledge to be perfect, and scientific investigation to be finished. 
Every considerable advance in knowledge would overturn the entire 
structure. 
The great mathematician and philosopher, Leibnitz, devoted much 
thought to what he called a Specieuse generate —an artificial language 
which would serve not merely for communication, but also as an aid 
in reasoning and invention. He died without publishing even the out¬ 
lines of his scheme; but from the success of the notation in which he 
clothed the ideas of the differential calculus we may infer that the 
non-developmeut of his scheme has been a loss to exact science. 
The inventor of Volapuk, J. M. Schleyer, of Constance, does not 
build on a philosophical classification ; he is pre-eminently a linguist, 
and builds his artificial language on a comparative study of the im¬ 
portant natural languages of the globe. The scheme comprises a dic¬ 
tionary of stems, taken mostly from the English language, and a gram¬ 
mar containing universally applicable rules for forming compound 
words, the cases of the noun, and a multitude of forms of the verb. 
He builds on a linguistic basis, and takes as his models the old syn¬ 
thetic languages. 
In this paper I propose to show that the proper basis for an artificial 
language is scientific analysis and classification ; and the two speci¬ 
mens of language so constructed will exhibit the enormous complexity 
of the problem. 
The inventor of Volapuk had before him the universally accepted 
notation for numbers, and in a language intended to be universal we 
naturally expect to find the nomenclature for numbers based on that 
notation. Butin Volapuk this expectation is not realized. The words 
for the digits are; 
