Wednesday, August 23, 2017

Explorations in bilingual metrics - Ukrainian speakers in Canada

Looking back twenty years at Budzhak-Jones & Poplack’s 1997 linguistic analysis 
of two generations of Ukrainian speakers in Canada

My great-grandfather’s cousin’s family back in Kiev
may all drive a "machine,"
but the rest of us here in Canada
will always drive a "car"


Svitlana Budzhak-Jones and Shana Poplack assembled a 10-hour corpus of spontaneous conversation from 7 local speakers from the Ukrainian expatriate community in Ottawa. The target speech was the speakers notion of Ukrainian, with English interspersions as they spontaneously arose. The authors call these multi-word code-switches “stretches (p.241). The authors documented their analyses and conclusions on several fronts in a 1997 article in the Journal of Sociolinguistics, Two generations, two strategies: The fate of bare English-origin nous in Ukrainian.” In the following critique of their assumptions, methods, and conclusions, I refer to the authors and their article as BJ&P.  BJ&P set up binary conditions to categorize loanwords from English into the Ukrainian spoken in Canada, as I've indicated in the preceding diagram.

Summary of B-J&P data and method
          BJ&P divided their set of speakers into two cohorts based on year of birth, then documented how well conversations matched target values for Standard Ukrainian as defined with the help of a native speaker (i.e., the primary author) and dictionary references from the previous 35 years. Each age cohort included speakers matched for place of birth, native language, language community, and years speaking English and Ukrainian. Facility measures included meta-linguisticdiscourse flagging (p.238) and grammaticality judgments from the primary author, supplemented with prescriptive formulae. BJ&P examined a primary conflict area and two conflicts within it, as well as a second conflict related to those.
(1) Inflectional markers on nouns
(a) Nouns from the Ukrainian lexicon nouns
(b)  Nouns borrowed” from the English lexicon
(2) Generational differences, as characterized by the birth year cohort of speakers
Their innovative approach for (2) was to compare the two cohorts to each other rather than to Standard Ukrainian. Thus, their baseline was close to 80% for correct inflectional productions rather than 100%. The higher scores came from the cohort born between 1920 and 1940. The cohort born between 1966 and 1977 performed at lower facility across all parts of BJ&P’s metric (1).1 At the time of the study, the speakers in each cohort fell into these age ranges: 25.6 +/- 5.6 and 66.6 +/- 9.62


Conclusions from B-J&P
B-J&P state their motivation for this area of research in their final paragraph (p.253): to establish the need for a metric of bilingualism. They propose that these data could support a metric based on comparing relative occurrences of inflectional endings. Truly bilingual” speakers perform better at reaching putative inflectional targets.

 --------------------------

1 These are not the actual birth years: I subtracted the reported ages from 1997, which was the publication date.
2 Again, this is not explicit in the article. I have used the age ranges that the authors give in order to compute a
median within the range.


 -------------------------------------------------------------------------------------------
Critique
I just referred to the "inflectional targets" as putative because that contention forms the majority of my critique. Before Monday- morning quarterbacking, let us review the innovative reach of  BJ&Ps work in its attempt to tease out how an individual in a dual language contact world separates, fuses, and switches between L1 and L2 when living in an L2 environment.
BJ&P compared two languages with differing morphological typology. English is analytic because it relies on word order and has few inflectional morphemes; Ukrainian is fusional as it loads its morphemes with more than one inflectional purpose. Specifically, BJ&P use evidence from how different generations of bilinguals integrate loanwords. BJ&P want to learn whether the words are integrated lexical additions or code-switching tokens. This underlies the question of how to categorize the multiplicative products of language contact. Is there a Canadian version of Ukrainian or is there a Ukrainian version of Canadian English? At what point is there a synthesis of Ukrainian and English?  The synthesis would have the syntax of the substrate language (Ukrainian) with additions from the superstrate lexicon (English).
Interestingly enough, the Canadian example is the obverse of the phenomenon of prestige considerations in thelanguage contact paradigm between an indigenous substrate and superstrate intrusive language. Here we have a substrate intrusive language (Western rural Ukrainian) and a superstrate majority language (Canadian English). The strata state of Canadian Ukrainian is unaddressed. There is some reference to divergent forms of Ukrainian. There are also early public school instruction materials in Ukrainian, used and written in Canada. They provide evidence of a language community separate from the European one.
BJ&P do not consider whether a speaker may want to mark an English word as
[+foreign] by retaining the English pronunciation.3
BJ&Ps work does not touch on these factors, which have become current areas of interest: heritage speakers; language identity; Canadian Ukrainian as a dialect; and language priming that takes place depending on the language setting. BJ&P relied on one of the paper’s authors for grammaticality judgments because she spoke European Ukrainian, but this might skew their ideas of grammaticality and standardization.
           BJ&P did not consider using other years for boundaries to define the Ukrainian waves of immigration.  Enumerating the birth years provides a new window into whether the speakers were speaking a dialect of Canadian Ukrainian related to their immigration wave. Wave I was pre-1914 (Duravetz 1988) and Wave II followed that and was prior to WWII (Makuch). The older cohort in the study may be from Waves I and II, and not necessarily Wave III. The heritage speakers may have learnt an evolving form of Canadian Ukrainian.

 -----------------------------------------------
3 For example, a monolingual English speaker could pronounce Picasso in a Spanish style (using / i /) rather than the standard / ɪ̈ / in order to emphasize an affinity for the Spanish-ness of the artist. Bilingual Spanish-English speakers would have more reason to express identity by code-switching” with just a twist of a vowel. On the other hand, they might not even have / ɪ̈ / handy in their phonetic inventory for a name they have always heard in Spanish. BJ&P wisely left out proper names from the study, presumably because pronunciation choices may have a more complicated interplay with a speakers wish to assert or create language identity.


-------------------------------------------------------------------------------------------------------------------------------------------------


The authors do not document the linguistic setting for interviews. The prime interviewer was a newly arrived, university-educated native speaker of Ukraine, who likely lacked the rural flavors of regional variation. University of  California researcher Jeff Elman (1977) found that the language environment in bilingual studies skews phonological boundaries (i.e. Voice Onset Timing or VOT) toward the primed language. That is, if the testing environment is Spanish speaking, then the bilinguals are more likely to categorize voiced and voiceless stops along the Spanish boundary of +4 millisecond rather than use the English-speakers +30 msec. boundary. Specifically, informants would hear "ball" for "Paul" if the researcher greets them and explains the task in Spanish.

Paradigms for setting up research on bilingual populations have evolved since 1997, so these detailed points do not disavow the original findings. The basis of any metric is its repeatability no matter the investigator. Refinements to a protocol are part of the research process. Addressing these critical points will move to a standard protocol, so future researchers can derive a consistent model that supports a metric for bilingual assessment. 
         

----------------------------------------------------------------------------------------------------


Duravetz, G. (1988, Sept), Ukrainian immigration to Canada 1891-1914. The Ukrainian  Canadian.  Retrieved from Virtual Museum of Canada (VMC) Web site: http://www.virtualmuseum.ca/sgc-cms/histoires_de_chez_nous-community_memories/pm_v2.php?id=story_line&lg=English&fl=0&ex=464&sl=5504&pos=1
Budzhak-Jones, Sand Poplack, S(1997), Two generations, two strategies: The fate of bare
            English–origin nouns in Ukrainian. Journal of Sociolinguistics, 1 (2), 225–258. DOI: 10.1111/1467-9481.00013

Elman, J. L., Diehl, R. L., & Buchwald, S. E. (1977). Perceptual switching in bilinguals. Journal of the Acoustical Society of America62(4), 971-974. DOI: 10.1121/1.381591

Makuch  A. (n.d.). The history of Ukrainians in Canada. Retrieved from Toronto Ukrainian Genealogy Group Web site: https://web.archive.org/web/20120131202612/http://www.torugg.org/History/history_of_ukrainians_in_canada.html  Note: Earlier archived versions of this article do not show an author.

-----------------------------
by Michelle K. Gross, postgraduate student assignment for Professor Minglang Zhou
SLAA629A Special Topics in Sociolinguistics -- Variationist Sociolinguistics, Professor Steven Ross
2015-October-16
This squib is archived as https://archive.is/wip/Bl8ua and as