¹ Lg. 31.31–5 (1955).

² The latter difficulty is of course explicitly pointed out (34) by Grimes himself, and I do not mean to belabor him. But the orthography he has chosen here forces his hand, and the question of decision is automatically raised. From Grimes's observation (32) that within a phrase Class 1 may never give way to Class 2 (though the converse does happen), it is also true that meqi- can scarcely be considered to belong to Class 1. But again, Grime's orthography does not cover this aspect.

³ G. L. Trager, article on Language, Encyclopaedia britannica 1955. The term is introduced in the section subheaded Communication.

⁴ Since submitting this paper for publication, I have read, with pleasure and profit, the article by Stockwell, Bowen, and Siiva-Fuenzalida: Spanish juncture and intonation, Lg. 32.641–65 (1956). In this generally excellent description, the authors give a programmatic treatment (641, 662–5) of certain vocalizations, or of certain systems or parts of systems of vocalizations; this is most welcome, particularly in view of the novelty of the field. But I object strenuously to the view implied by one sentence (662): 'The following pairs occur in Spanish.' In principle, a given system may 'co-occur with' one type of Spanish, or even with all of Spanish (though I doubt it) ; if a set of systems should turn out to be in fact coextensive with Spanish, this would be, from an analytic viewpoint, by accident and by exclusion (i.e. after all analogous geographically and socially neighboring systems had been surveyed and classified). An example, formulated impressionistic ally: the Albanian and Greek speakers of southern Italy have vocalizations that are ‘South Italian’ (a tentative fiction) ; the Albanians of southern Greece have ‘Greek’ or ‘South Balkan’ vocalizations—as have many though probably not all Tosk Albanians of Albania; many Gegs have vocalizations in common with speakers of Serbian; and I know at least a few immigrant Americans who speak Albanian accompanied by subtypes of ‘American’ vocalizations. It is important to observe this distinction—especially since for the first time formal linkages between linguistic and other cultural subsystems are being stated with some degree of rigor and elaboration. An unfortunate precedent should be avoided.