40 COMPARATIVE LANGUAGE 



some kind of formal distinction between singular and plural is com- 

 mon to practically all languages, but some have also a special form 

 for the dual, which is the linguistic expression of association between 

 objects occurring in pairs. Had familiar objects occurring in sets of 

 five, like the fingers and toes, been as numerous as those occurring in 

 pairs, the hands, feet, ears, eyes, etc., their association with one 

 another might equally well have reflected itself in another gram- 

 matical category. Not that we are to imagine any conscious effort 

 in the beginning to differentiate objects occurring in pairs and to 

 provide their names with endings significant of this. It is rather 

 that, given an expression, let us say, for " the hands," not in itself 

 indicative of their number, the expression for "feet," "eyes," etc., 

 whether in their initial creation or later, would be assimilated to this, 

 until finally from a sufficient number of such forms there would 

 arise a consciousness of the significance of the common element, 

 which now becomes a "dual ending." But this consciousness of the 

 significance of the dual is only the prelude to its gradual loss as 

 a distinct formal category. For with the increasing clearness in the 

 perception of relations, the difference between one object and more 

 than one comes to be felt as the all-important one and the dual is 

 sooner or later merged in the plural. 



The vocabulary is also significant of modes of thinking. It has 

 often been noted that people on a low stage of civilization show 

 what seems a high degree of differentiation, as when they have 

 separate words for washing, according as they mean washing the 

 hands, washing the face, etc. But in reality this is only a lack of 

 generalization, characteristic of what is termed fragmentary thinking. 

 The savage does not differentiate the concept wash into wash the 

 hands, wash the face, etc., but the notions of washing the hands, the 

 face, etc., are distinct, concrete concepts, not yet put into relation 

 with one another and generalized under the abstract wash. 



But aside from the psychological significance of such general 

 linguistic phenomena, the every-day problems of the comparative 

 grammarian in the narrower sense are, to a large degree, psycho- 

 logical. For whether he is dealing with forms or with syntax, he 

 finds that the history of the individual word or construction is 

 affected by its associations. Changes in the form of a word are by 

 no means confined to those caused by the regular phonetic processes, 

 but are frequently due to the influence of forms which are for any 

 reason associated with it in the mind. All the phenomena classed 

 under Analogy, Leveling, or Contamination are examples of associa- 

 tive interference. If the child says teached for taught, if blowed for 

 blew is not uncommon, and if we all now say snowed for an earlier 

 snew, it is owing to the influence of the great body of words in 



