A loanword is a word that is adopted or borrowed from another language with its articulation, orthography, and inflection adapted for the recipient language [1, pp. 396]. We (@stephsamson, @ketxd) examine the question of whether there is a relationship between the number of loanwords in a language and its speaker size. Using loanword percentage from the World Loanword Database and speaker sizes from Ethnologue, we calculated the correlation coeffcient to see if there is such a relationship. We did not find a considerably strong correlation, whether positive or negative, between the two variables but did find slightly stronger correlations when considering size of L2 speakers. (i.e., people who speak the recipient language as a second language).
The database used is the World Loanword Database (WOLD) borne out of Haspelmath’s and Tadmor’s collaborative efforts. WOLD has 41 languages. Of the 41, one is Old High German which has zero living native speakers and was omitted from the set of data points since this project is not concerned with extinct languages. For each language in WOLD, an expert in the said language contributed to its vocabulary [3].
In WOLD, there are 5 degrees of borrowing certainty:
-
Clearly borrowed
-
Probably borrowed
-
Perhaps borrowed
-
Very little evidence for borrowing
-
No evidence for borrowing
Loanword percentages in WOLD only consider level-1 and level-2 loanwords [3, pp.12-13].
The survey of loanwords are done on the basis of a fixed list of 1,460 lexical meanings [2, pp.1,5]. It should be noted that in some languages, these meanings may have more than one word representation or none at all. The sample of 40 languages (Old High German excluded) includes "languages indigenous to all continents and belonging to many language families" [2, pp.3]. Of greater interest to this project is that the speaker size of this language sample is in many orders of magnitude: from the hundreds (Ceq Wong) to the hundreds of millions (English, Mandarin Chinese) [2, pp.3].
However, the sample also has its shortcomings primarily due to the lack of experts who would volunteer to curate vocabularies for WOLD. This led to "over- or under-representation" in regions or language families [2, pp.3]. There is also a "skewage in statistics": if there is a word borrowed into parent language P and it is inherited by descendant languages A, B, and C, it would be counted as three loanwords when it actually represents a single borrowing event [2, pp.3].
The number of speakers are derived from Ethnologue.
[1] H. Bußmann. Lexikon der Sprachwissenschaft. Alfred Kröner Verlag Stuttgart, 2008.
[2] M. Haspelmath and U. Tadmor. Loanwords in the World’s Languages: A Comparative Handbook. De Gruyter Mouton. De Gruyter Mouton, 2009.
[3] M. Haspelmath and U. Tadmor, editors. WOLD. Max Planck Institute for Evolutionary Anthropology, Leipzig, 2009. Accessed: 12 Jul 2017.