Dear Monika, 

I completely understand (and partly share) your frustration. The distinction between 'early' versus 'late' L2 learners and the (somewhat different !) distinction between 'simultaneous' and 'sequential 'bilinguals are linked to the concept of 'critical periods' or 'sensitive periods' in language acquisition (Penfield/Roberts; Lenneberg, 1967).

These days, most proponents of the critical period hypothesis (CPH) in L2 acquisition would likely agree that the relevant 'window of opportunity' (which might allow you to reach native-like proficiency) strongly depends on the linguistic domain (early for phonology, later for syntax; potentially absent for lexical and conceptual semantics). From this perspective, clinical studies of the kind mentioned in your question would definitely have to specify which domain they are interested in. But even within such domains, there are subdomains (e.g., segmental vs. supra-segmental phonology, logical versus conceptual semantics, etc) that may have distinct temporal dynamics and distinct windows of opportunity. And even within such subdomains (e.g., within "inflectional morphosyntax"), acquisition of one particular feature or structure may typically happen earlier than another one (and be dramatically influenced by the similarities and differences between L1 and L2).

Crucially, however, more recently researchers in the field of neuro-bilingualism have questioned whether the concept of a 'critical period' makes much sense at all for second language acquisition (SLA). While hardly anyone doubts that, as a rule of thumb, 'the earlier, the better in SLA', whether there are CPs and clear cut-off points for larger linguistic domains such as phonology or syntax is very controversial. If not, then using dichotomies such as 'early' vs 'late' in studies (e.g. using ANOVAs) seems highly questionable, and instead one might want to use models dealing with continuous variables (regressions, mixed models, etc.). My own research (primarily using event-related brain potentials, ERPs) in various linguistic domains (and clearly focusing more on L2 than L1 acquisition) has led me to doubt the usefulness of clear cut-off points. Based on inspiring discussions with other colleagues around the world (often working in other fields of L1 or L2 acquisition), my impression is that many researchers in the field are losing confidence in traditional cut-off points (which, at least sometimes, seem to be based on questionable assumptions). 

In sum, I'm afraid there is no simple answer to your question, but you might have to  prepare for an approach that does not rely on easy-to-handle cut-off points ...

I've not included references in my answer above. If you need a starting point, perhaps have a look at my RG website or this debate: https://www.researchgate.net/post/Do_you_think_that_there_is_a_critical_period_for_language_acquisition_or_not

(My reply to that question has an article attached to it.)

I hope this helps. Given how controversial the field is, I'm sure there are other researchers who would be able to provide you with a more encouraging perspective re cut-off point .... :-)

Cheers, Karsten

Hi Mahmood,

Great thanks for the response.

I am still trying to figure out what the cutoff could be based on the neurodevelopmental literature. 

Two recent reports using resting state fMRI have shown differences in functional connectivity between children who started acquiring their L2 at age 3 vs. children who started acquiring their L2 at age 6.  The cutoff they seemed to use was age 5.

Here are the links to the reports:

Liu et al. 2017: http://www.sciencedirect.com/science/article/pii/S0093934X16302097?via%3Dihub

Berken et al. 2016: https://www.ncbi.nlm.nih.gov/pubmed/26818505

Cheers,

Monika

Dear Monika, 

I completely understand (and partly share) your frustration. The distinction between 'early' versus 'late' L2 learners and the (somewhat different !) distinction between 'simultaneous' and 'sequential 'bilinguals are linked to the concept of 'critical periods' or 'sensitive periods' in language acquisition (Penfield/Roberts; Lenneberg, 1967).

These days, most proponents of the critical period hypothesis (CPH) in L2 acquisition would likely agree that the relevant 'window of opportunity' (which might allow you to reach native-like proficiency) strongly depends on the linguistic domain (early for phonology, later for syntax; potentially absent for lexical and conceptual semantics). From this perspective, clinical studies of the kind mentioned in your question would definitely have to specify which domain they are interested in. But even within such domains, there are subdomains (e.g., segmental vs. supra-segmental phonology, logical versus conceptual semantics, etc) that may have distinct temporal dynamics and distinct windows of opportunity. And even within such subdomains (e.g., within "inflectional morphosyntax"), acquisition of one particular feature or structure may typically happen earlier than another one (and be dramatically influenced by the similarities and differences between L1 and L2).

Crucially, however, more recently researchers in the field of neuro-bilingualism have questioned whether the concept of a 'critical period' makes much sense at all for second language acquisition (SLA). While hardly anyone doubts that, as a rule of thumb, 'the earlier, the better in SLA', whether there are CPs and clear cut-off points for larger linguistic domains such as phonology or syntax is very controversial. If not, then using dichotomies such as 'early' vs 'late' in studies (e.g. using ANOVAs) seems highly questionable, and instead one might want to use models dealing with continuous variables (regressions, mixed models, etc.). My own research (primarily using event-related brain potentials, ERPs) in various linguistic domains (and clearly focusing more on L2 than L1 acquisition) has led me to doubt the usefulness of clear cut-off points. Based on inspiring discussions with other colleagues around the world (often working in other fields of L1 or L2 acquisition), my impression is that many researchers in the field are losing confidence in traditional cut-off points (which, at least sometimes, seem to be based on questionable assumptions). 

In sum, I'm afraid there is no simple answer to your question, but you might have to  prepare for an approach that does not rely on easy-to-handle cut-off points ...

I've not included references in my answer above. If you need a starting point, perhaps have a look at my RG website or this debate: https://www.researchgate.net/post/Do_you_think_that_there_is_a_critical_period_for_language_acquisition_or_not

(My reply to that question has an article attached to it.)

I hope this helps. Given how controversial the field is, I'm sure there are other researchers who would be able to provide you with a more encouraging perspective re cut-off point .... :-)

Cheers, Karsten

Dear Karsten,

Thank you so much for the elaborate answer and the link to the other question/discussion.

You have brought up several important points that I very much agree with. As you said, making a single cutoff for different linguistic domains will most likely be tricky since the cutoff is an artificial concept and it may thus be of little scientific use. I think applying separate cutoffs for different domains might be a safer (yet still not ideal) approach in examining neurobiological underpinnings of earlier vs. later language acquisition. 

In the past few years I have worked with two clinical populations - patients with tumors and epilepsy. Those patients typically undergo lexico-semantic language evaluations as part of their clinical language mapping. I believe I should focus on choosing the early-late cutoff based on the literature on lexico-semantic neurodevelopment. I believe age 5 as the cutoff might work but I will certainly have to investigate this further. 

I appreciate your help!

Best wishes,

Monika

Hi again,  yep, I completely understand. Re lexico-semantic development, a student in our dept recently defended her PhD thesis on depth (rather than breadth) of lexical representations in bilingual versus monolingual kids (also taking AoA into account), and we arrived at the same kind of debate: what is a good cut-off point for L2?  Once again, she had started with a cut-off point based on the literature (something like 3 years, if I'm not mistaken), but then had to admit that her research didn't support it at all. Her conclusion: perhaps using age of (school) entry (because writing skills have an effect) or, probably even better, use age as a continuous variable. The student (Myrto Brandeker) has been working with my colleague Elin Thordardottir (in case you want to check their research out). Good luck w/ everything ! (Your research on pre-surgery language mapping looks interesting too.) Cheers, Karsten

How interesting! Thank you for sharing the conclusion from your colleagues' research. I will check their work. 

Good luck with everything as well!

Cheers,

Monika