Penn Linguistic Conference


Anecdotal evidence in second language acquisition and its implications

Anecdotal evidence in second language acquisition and its implications

The primary purpose of the theory of human language, grammar is to be able to tell well-formed sentences from those not well-formed: this has been the working hypothesis in generative grammar ever since Chomsky (1955). The very minimal requirement of the grammar is that it be compatible with the observable facts about language. Same is true for language acquisition: theories should be fully compatible with whatever observational data comes their way, as well as with the data obtained in controlled conditions. One variety of data which is not obtained in the lab but is observed is anecdotal data. It is the value and the role of this data that we would like to highlight presently.

Anecdotal evidence and case studies have been repeatedly shunned from language acquisition inquiry (with only a few notable exceptions of case studies, e.g., Lardiere 2007), yet the reasons for this are rarely made clear. In this paper, we argue that anecdotal evidence actually presents researchers with an invaluable tool to inform their hypotheses while not necessarily serving as a great tool to test those hypotheses in a controlled setting. We investigate three historical cases which pose nontrivial questions to some modern theories of second language acquisition generally and the critical/ sensitive period hypotheses in particular. The cases are:

After careful analysis of the documented evidence (see below), we consider the manner in which each of these cases challenges modern theories of second language acquisition. Relevant background for these cases is given in Table 1 below. Our main argument is that all these cases point to inviability of the fundamental difference hypothesis (FDH) and related hypotheses of critical period/sensitive periods as well as the thesis on impossibility of ultimate attainment – all of which resurface every decade since Lenneberg (1967) in one incarnation or another (cf. Abrahamsson and Hyltenstam (2009), Bley-Vroman (2009), and Meisel (2011) and also see an overview of critical/sensitive period(s) in Wang (2018)). The central claims we lay are as given below.


Age group

Previous language(s)

Language acquired

Duration of exposure

Type of evidence

N.N. Miklouho-


Russian, French, German, Spanish

Bongu, incomplete

~1.5 year


J. Manjirō




~5 years

oral account, letters

W. Buckley




~20 years

oral account

Table 1. Relevant language background for the cases considered

The Miklukho-Maklay (MM) case. Although the acquisition of Bongu was incomplete and the time of exposure was limited, MM was able to communicate with natives with a significant degree of fluency, as evidenced by his notes in the diary. Further, the collected by WW items of phraseology and distinct lexical items let us gauge the depth of MM’s command of the language (subjunctive, imperative moods; wide range of tenses, etc.). Given the limited time of his exposure (about one year) and additional – albeit less significant – contact with over 13 other aboriginal languages, such rapid acquisition runs decidedly counter against “sensitive period” for SLA. It up to speculation, however, whether MM managed to reach nativelike proficiency, to which his diaries do not attest. Account of William Buckley, however, does attest to such proficiency, hence our second case [1].

The William Buckley case. Buckley was at least 23 years old when he left for Australia as a convict (past the last “sensitive” period of 17 years). He escaped and chanced upon Wathaurong people who he lived with for about 20 subsequent years, becoming the head of their tribe. Notably, Buckley recollects the silent period in his acquisition, and points to the fact that just after two years of exposure he acquired nativelike fluency of a language drastically different from his native English. Again, the rapidity of acquisition under these circumstances – which can be corroborated by socio-ethnical considerations (natives only treated him as “one of them” upon ultimate attainment) – is not accounted for by the critical/sensitive periods or fundamental difference hypothesis which do not allow ultimate attainment per se. Our third case, however, offers ambiguous evidence.

The Manjirō case. Manjirō, a Japanese boy of 14 years (end of the last sensitive period), was shipwrecked and saved by an American whaling ship. Manjirō was the youngest of the five companions picked up by the Americans and over the period of 11 months (Jan-Nov) he managed to acquire the language to the degree that he was able to work with the crew and communicate with the captain with marked fluency. Notably, Manjirō remarks that the oldest of their crew was much less successful in acquisition of English. We look into the factors which could have contributed to this in our analysis. Further, Manjirō’s letters from his time in the US are available and can be scrutinized to see a clear trajectory of acquisition. Once again, the rapidity of the acquisition points to Manjirō’s level after a year of exposure being compatible to that of MM’s. This as well runs counter to the sensitive periods and the fundamental difference hypothesis.

Concluding, it is important to note that naturalistic language acquisition[2] should have been one of the first places to look for mechanisms of language acquisition unblemished with external variables, but very few studies did so (save early studies like Felix (1981), Pica (1983)). Instead, based on decidedly limited evidence obtained in the classroom or in a laboratory setting, a number of theories which are directly incompatible with evidence in plain sight were conceived. While our elaboration does not present resolute data against FDH, etc., we hope that it lends itself for didactical, instructive purpose and demonstrates quite plainly that hypotheses and theories of language acquisition – just any other hypotheses and theories – can and should be informed by the available anecdotal evidence. Compatibility with anecdotal evidence, in turn, can serve as a litmus test for any hypothesis.


[1] MM’s diaries also mentioned an aboriginal boy of 14 years who was fluent in Russian after 4 months on a ship with MM. We consider this case in passing only

[2] Importantly, not “immersive” instruction in a classroom which does not constitute naturalistic acquisition.