Submit to this Journal Review for this Journal Propose a Special Issue

Article Menu

Printed Edition

A printed edition of this Special Issue is available at MDPI Books....

Share Help Cite Discuss in SciProfiles

Open AccessArticle

Peer-Review Record

Is There an Effect of Diglossia on Executive Functions? An Investigation among Adult Diglossic Speakers of Arabic

Languages 2022, 7(4), 312; https://doi.org/10.3390/languages7040312

by Najla Alrwaita¹, Lotte Meteyard¹

, Carmel Houston-Price¹

and Christos Pliatsikas^1,2,*

Reviewer 1:

Thaís De Sá

Reviewer 2: Anonymous

Reviewer 3: Anonymous

Languages 2022, 7(4), 312; https://doi.org/10.3390/languages7040312

Submission received: 1 April 2022 / Revised: 25 October 2022 / Accepted: 28 November 2022 / Published: 16 December 2022

(This article belongs to the Special Issue Multilingualism: Consequences for the Brain and Mind)

Round 1

Reviewer 1 Report

The paper “Is There an Effect of Diglossia on Executive Functions? An Investigation on Young Adult diglossic Speakers of Arabic” is one of many investigations about distinctions between monolingual and bilingual cognitions. The thing that distinguishes this paper from others is that the investigation occured in arabic diglossic contexts, an original aspect to be considered. From an Adaptive Control Hypothesis perspective, authors hypothesized that if there were limited opportunities for language switching to occur, as in Arabic diglossia, effects on cognition (on cognitive tasks assessing the executive function domains of inhibition, switching and updating) should be more expected than in contexts where code switching is more legitimated. The article presents a good literature review, which is relevant to the topic; and confirms the relevance of studying effects on cognition in diglossic context. I wish there was more about the grammar and the syntactic distinctions of Modern Standard Arabic (MSA) and the colloquial dialects for non-Arabic speakers. This would help the argument that diglossia also needs to be studied and it could clarify more the found results. The arguments and discussion of findings are coherent, balanced and compelling. It is a good paper and it should be published.

Author Response

We would like to thank the reviewer for drawing our attention to this. In response to this comment, we have added the following sentences to our text (Page 5), as well as a relevant reference:

“These two varieties also have different levels of language complexity. For example, H includes an elaborate system of inflections, such as: number, gender, person, case and definiteness. This results in a more complex grammatical phrases in H compared to L. L, on the other hand, is known to have a simplified morphological system which lacks inflections (Albirini, 2016). From a syntactic perspective, H is known to have a SVO word order while L has an SVO order (Boudelaa & Marslen-Wilson, 2013)”

Reviewer 2 Report

This article presents an empirical study on the potential advantages of diglossia when it comes to (non-verbal) executive function tasks. The idea of investigating this is interesting and I’m sure that there will be a large group of readers interested in this topic. Nevertheless, I have some serious concerns about the comparisons presented here. The results and analyses are discussed at length, so much so that it confuses the reader and there is a lot of redundant information included. I have some doubts about the analyses, but find it difficult to judge at the moment due to the confusing structuring of information. However, I think the main problem with this article is the comparison between two groups from completely different backgrounds and cultures. Although the authors do (partly) acknowledge this in the discussion, it does not make up for the fact that this is an ‘apples and oranges’ situation. Instead, I recommend the authors to explore if any factors measured within the diglossic group enhance executive functions.

Detailed comments per section:

Introduction

Diglossia is introduced and Colloquial Arabic/MSA given as an example and argued to be an ideal candidate, but of course within Arabic too there is great variety in the distance of the colloquial language from MSA, the everyday use of both varieties, etc.

Some more could be said about the social and political situation concerning the use of different types of Arabic, for example with regards to everyday use. Please paint a picture for readers who are less familiar with this part of the world.

Actually, there is quite a large body of literature reporting bilingual advantages. The two references do not cover this and it is not clear why the authors choose to give so much more weight to studies reporting null findings. Please also check more recent work by the Bialystok group as well as other groups. Some suggested references are below; references therein may also be useful. If the authors argue that indeed in young adults there are no bilingual advantages or they are difficult to find, why would diglossia be any different? Wouldn’t any effect be expected to be stronger in a Dual Language context? So why then test diglossia?

Bialystok, Ellen. 2011. Reshaping the Mind: The Benefits of Bilingualism. Canadian Journal of Experimental Psychology 65: 229–35.

Costa, Albert, Mireia Hernández, and Núria Sebastián-Gallés. 2008. Bilingualism aids conflict resolution: Evidence from the ANT task. Cognition 106: 59–86.

Soveri, Anna, Matti Laine, Heikki Hämäläinen, and Kenneth Hugdahl. 2011. Bilingual advantage in attentional control: Evidence from the forced-attention dichotic listening paradigm. Bilingualism: Language and Cognition 14: 371–78.

Treffers-Daller J, Ongun Z, Hofweber J and Korenar M (2020) Explaining Individual Differences in Executive Functions Performance in Multilinguals: The Impact of Code-Switching and Alternating Between Multicultural Identity Styles. Front. Psychol. 11:561088.

Methods

I initially misinterpreted the term ‘young adult’ as this is generally taken to mean around the age of 18-22, but the article reports data from participants with a mean age of 30 (rounded, for the Arabic-speaking group). These are just adults. This should be made clear much earlier on in the acticle.

In addition, the English monolingual group is quite a bit younger. Could this confound the results? This is important as the authors themselves argue that young adults (i.e. the monolingual group in this case) may be cognitively at peak performance.

There is another issue with the comparison: the two groups were not only from different language backgrounds, but also from completely different cultural and social backgrounds (including education, which differs vastly between different countries, and their studies, SES(?) ).

Standard Arabic vocabulary test: what were participants’ scores? What was the exclusion threshold?

Results

How were the cutoffs for data filtering/cleaning for each task determined? They seem rather arbitrary and there is no justification provided.

How were the levels of the fixed factors coded? Was treatment coding used?

If model convergence is difficult due to the random effects, the authors could try to remove their correlations (see https://cran.r-project.org/web/packages/lme4/vignettes/lmer.pdf, double-bar notation)

Providing an overview of the accuracy results (in a figure or table) would help. At present, the reader has to dig through many paragraphs of text to find them.

Page 9 (Flanker accuracy): “The inclusion of group did not contribute significantly to the overall model fit […]. Further analysis showed no significant difference between Diglossics […] and Monolinguals […].” Which further analyses were these, if the factor was not included in the models? This statement is repeated for the interaction.

In general, the description of the model output is very odd for a linguistics article. I suggest rewriting the Results section so that 1) there is a general header ‘Analyses’ that describes how the models were built (could be under Methods already) 2) there is only a description of the final models in the results section 3) models without convergence or with singularity warnings are discarded 4) numerical results that are provided in the tables are not repeated in the main text 5) Factors/variables that are not included in the models are not discussed as if they are. Illustrating this last point, the statement “However, The interaction between group and condition did not significantly contribute to the overall model fit (χ2(2)=3.34,p=0.1882). There was a significant interaction between group and condition when comparing Incongruent and Neutral, where Monolinguals showed a smaller difference between the two conditions (β=0.022,SE=0.003,t(782.9)=5.74,p=0.0001).” is very confusing: if there is no interaction, then why is there an interaction discussed?

Table 1: the caption states that missing comparisons were obtained through releveling (presumably within lmer), whereas elsewhere in the article it states that these were obtained with emmeans. Which is it?

Since all tasks were used to test the same hypothesis, were corrections for multiple comparisons applied?

I was very disappointed when I found out that the authors had only done between-group comparisons and didn’t attempt to discover if any within-group factors affect executive functions. For example, the authors discuss the effects of dominance, domain of use, and formal education in the introduction. In addition, they ran the LSBQ and a vocabulary test but don’t seem to do much with these data. I think this warrants further investigations.

Discussion

It is difficult to interpret the results meaningfully due to the differences between the experimental and the control group. It is therefore almost ironic that the authors use unmatched external factors such as SES as an argument for why other studies have found a bilingualism effect; I could very easily use the same argument here for why the current study found no such effect.

Clarity of writing

Some spaces are missing (this may have happened in the submission process or before), such as in abilingual, tolanguage, and includeexecutive on page 1.

On page 8 under Results: Kendrick et al. (2019) are not the authors of the lme4 package, please use the appropriate reference here.

Author Response

Detailed comments per section

Introduction

- We thank the reviewer for this comment. Diglossia in the Arab world is indeed a multifaceted phenomenon, which we chose not to describe in full detail, in order to avoid deviating from the main focus of the paper. Nevertheless, we have now added the following text to our manuscript (Page 5), alongside some additional references:

“It is worth noting that Arabic diglossia takes different forms in different countries, and several factors affect the local Arabic variety that is spoken. For example, Rosenhouse and Goral (2006) divided Arabic dialects on the basis of four categories: First, geographical dimension, separating Eastern dialects, those spoken in Saudi Arabia, Yemen, Kuwait, Oman and the United Arab Emirates, from Western dialects, including those spoken in Morocco, Algeria, Tunisia, and Libya; second, sociolinguistic background, including Bedouin vs. al-Hadar backgrounds; third, religious background (Islam, Christianity or Judaism); and last, gender, age and education (Alsahafi, 2016). It is important to note that mutual intelligibility often exist between local varieties, and it has been suggested that the more geographically close countries, the more intelligibility is found between the speakers of these countries’ varieties. It has also been reported that Arabic speaking individuals often modify their variety to approach another individual’s variety (Giles, 1973; S’hiri, 2002) . This usually happens during intralingual contact and the process is referred to as convergent accommodation (Alsahafi, 2016). For instance, dialects of Western Arab countries, such as Tunisia and Morocco, have diverged from Standard Arabic more noticeably than dialects of Eastern Arab countries, such as Jordanian or Palestinian. Thus, it is expected that Moroccans, for instance, would modify their dialect when speaking to Jordanians (Alsahafi, 2016)”.

Bialystok, Ellen. 2011. Reshaping the Mind: The Benefits of Bilingualism. Canadian Journal of Experimental Psychology 65: 229–35.

Costa, Albert, Mireia Hernández, and Núria Sebastián-Gallés. 2008. Bilingualism aids conflict resolution: Evidence from the ANT task. Cognition 106: 59–86.

We thank the reviewer for this comment. We intentionally chose to not elaborate much on the literature on the presence or not of bilingual advantages, as this has been covered extensively in a great number of other papers, and it was never our focus to provide a full overview of the findings. We recognise now that our presentation of the evidence may not appear balanced, so we have added some more references that provide evidence for bilingual advantages in younger adults.

Regarding studying diglossia and SLCs in general: As we mention in the text, most of the available literature on executive functions and bilingualism has not paid close attention to the particular linguistic context that bilinguals find themselves in, making it a real possibility that the published studies represent all types of contexts as defined by the ACH, despite the fact that each context provides different opportunities for code-switching, and therefore may have different effects on cognition. This manuscript, alongside another one that we recently published focusing on older adults, aims to test diglossia as a good example of the Single Language Context where individuals have limited opportunity for switching between languages/dialects, but still they face great needs for inhibition this was also fuelled by some recent finding from diglossics older adults, see below). If we observed particular effects for this context, we would be able to inform the design of future studies with respect to the linguistic context they use. If switching proves to be the essential for enhanced EFs this will also add to our understanding of the available literature on bilingualism.

Regarding studying younger adults: We don’t necessarily argue that bilingual benefits more or less difficult to be found in young adults- we are simply arguing that the mixed evidence comes from young groups with mixed characteristics/language use backgrounds. On the back of this we argue that, if anything, we need more evidence with designs that account for the characteristics of the language environment. SLCs are environments with very particular characteristics, including limited opportunities for code-switching but greater needs for inhibition, and therefore are appropriate for isolating effects on the respective domains. Indeed, in a recent study with older adults we reported effects of diglossia tasks tapping inhibition and switching. If this was to be replicated here, it would show that context is more crucial than age for these effects; the fact that it didn’t further corroborates the peak performance hypothesis, even for young diglossics. We have now refer to that study as follows (page 9-10):

“Some additional corroborating evidence also comes from a recent study comparing Arab diglossic older adults to bilinguals and monolinguals, which reported significant effects of diglossia, including a diglossic benefit in tasks tapping inhibitory control and switching (Alrwaita et al., revised & resubmitted). The authors interpreted their findings as the outcome of long-term experience in inhibiting the non-target language at SLC contexts.”.

We have also refined our predictions as follows (Page 12-13), and we have modified our discussion accordingly in order to reflect this dimension as well.

“To investigate whether Arabic diglossia has an effect on executive functions, we compared young Arab diglossics (in their early to late twenties on average) to English-speaking monolinguals in a series of tasks that have been reported to demonstrate effects of diglossia in older adults (Alrwaita et al., revised & resubmitted), namely the Flanker, Stroop, and Color-shape tasks. If our results replicate previous evidence for diglossic benefits (Alrwaita et al., revised & resubmitted; Antoniou & Spanoudis, 2020), this would suggest that diglossic situations, in general, enhance performance in tasks tapping executive functions, irrespective of the amount of language switching (which, as discussed, differs between Cyprus and the Arab world) or the age of the participants. Conversely, if we fail to report any effects of diglossia, and taking into account the findings from older Arab diglossics, this would constitute evidence for the peak performance hypothesis even for young diglossics in strict SLCs as they materialise in the Arab world.”

Methods

-These are very good points presented by the reviewer. The age range for our diglossic group was largely determined by the availability of participants during the pandemic, when this data was collected. Nevertheless, care was taken that no participants were included that could be characterised as “old”. It seems that there is no general agreement of the age range of “young adults”. For instance, Gold and colleagues (2013) tested monolingual young adults (mean age= 32.2), and bilingual young adults (mean age=31.6). We also acknowledge that much (but not all) of the discussion surrounding the peak performance hypothesis has emerged from samples in their early 20s We have mentioned in our discussion that both groups still lie within the gray area of young adults as defined by bilingual studies (Donnelly, Brooks, & Homer, 2019). At the end of the day, this remains a matter of labels. We have now included an indication of our age range in the “This study” section to avoid any confusion. We have also removed “young” from the title of the paper.

Furthermore, to avoid an effects related to age, including the average age difference between our groups, all our models in our current version of the manuscript included Age as a covariate. The pattern of results remained the same, suggesting that age does not interact with any effects of diglossia, at least for our current age range.

-We fully acknowledge this limitation, which arises by the fact that this is the best comparison we could provide, given that there are no education-matched non-diglossic young adults in Saudi Arabia, and also no diglossics in the UK. Although we recognise that educational systems may be considerably different between the two countries, we at least ensured that both groups had university-level education. Regarding SES, unfortunately we did not use a measure of it, and there may be a possibility that it significantly differs between the groups. However, the lack of between-group differences with respect to the effects that are traditionally used as evidence for bilingual advantages (e.g. the Flanker effect), actually suggests that SES may not be relevant such effects. The overall difference that we found (i.e. slower overall RTs for diglossics) may indeed relate to cultural or SES factors that we have already discussed in our Discussion. We have now also included a reference (Samuel et al., 2018) which attributes some between-groups differences in similar tasks to differences in cultural background.

Standard Arabic vocabulary test: what were participants’ scores? What was the exclusion threshold?

-We have added this information in the text as follows (Page 15):

“To estimate participant’s vocabulary size, measured as words known out of 50,000 words, all “yes” answers to real words are given a score of 500 to form an unadjusted vocabulary score, and each “yes” answer to a false word deducts 2500 points from the unadjusted score to form an adjusted vocabulary score. The final adjusted score gives the participants total vocabulary knowledge. Based on this, our diglossic group had an average of 87.59% (SD=12.17) of “yes” responses to real words, an average of 5.25% (SD= 3.69) to “yes” responses to non-words, and an estimated average of 30,672 (SD= 7,329; median= 30,375; range 15,000-42,000) known words. No one was excluded from the diglossic group based on this score, as they all demonstrated at least basic competency in standard Arabic (Masrai & Milton, 2019)”.

Results

How were the cutoffs for data filtering/cleaning for each task determined? They seem rather arbitrary and there is no justification provided.

-For data filtering, we have now clarified that we have removed data that was 2 standard deviations below or above the median. This should be effective at removing extreme values (for a given task), i.e. reaction times that are extremely fast and therefore not legitimate responses to the task, or reaction times that are extremely slow and therefore like to result from a loss of concentration on the trial.

How were the levels of the fixed factors coded? Was treatment coding used?

-We used the default treatment coding when running all models, and relevelled conditions to extract all possible comparisons between condition levels (i.e. congruent, incongruent, neutral conditions for the Stroop and Flanker tasks).

-We have re-run the analyses, including removal of correlations between random effects, for all models. We found that convergence issues persisted with the removal of correlations and where model simplification was needed, we had to remove all slopes and run integer-only models. Details are provided in the Results section and alongside results Tables.

Providing an overview of the accuracy results (in a figure or table) would help. At present, the reader has to dig through many paragraphs of text to find them.

-We have now re-run all the analyses to simplify the reporting of results, and substantially rewritten the results section. Maximal models were run initially, and then simplified if convergence issues arose. Details of model simplification are provided with results Tables, both for accuracy and reaction times. Rather than using both coefficient / model estimates and comparisons of model fit for interpretation of key results, we now only report coefficient estimates and interpret those. Comparisons of model fit are now only reported when reporting convergence issues (e.g. does a simplified model fit the data as well as a maximal model that had convergence issues?).

The following has been added under section 4.0 Results (page 18):

“Model fitting followed the same procedure for all data. Age was entered as a covariate. Task Condition (e.g. congruent, incongruent, neutral; stay or change trials) were entered as main effects. Bilingual Group (monolingual or diglossic) was entered as a main effect. We also tested the interaction between Group and Condition (i.e. whether monolingual or diglossic participants responded to the task conditions differently).

First, a maximal model was fit with random intercepts for participants and correlated random slopes for within-subject conditions. If there were convergence issues with the maximal model, the first step was removing correlations between random effects. If convergence issues persisted (as they did for all models that initially had convergence issues for the maximal model) we then removed random slopes and ran a model with participant intercepts only. We compared model fit for the maximal model and the final simplified model, and not any differences in the model output where the simplified model did not fit the data as well as the maximal model. Where a factor had more than two levels (Flanker and Stroop tasks with incongruent, congruent and neutral conditions) we relevelled the reference factor to extract all possible comparisons between conditions, and used emmeans (Lenth et al., 2021) to report the relevant comparisons for the interaction between group (diglossics vs monolinguals) and condition (incongruent, congruent, neutral).”

-We have used relevelling to get comparisons for main effects (e.g. incongruent, congruent and neutral level comparisons for a task). To interpret interactions for factors that have more than two levels, we used emmeans to report all relevant comparisons for that interaction (this applies only to results for Reaction Times for the Flanker and Stroop tasks, in which we evaluated the interaction of bilingual status (diglossic vs monolingual) and condition level (congruent, incongruent, neutral).

Since all tasks were used to test the same hypothesis, were corrections for multiple comparisons applied?

-We have not corrected for multiple comparisons, as the tasks are separate and the data arising from each task is separate. Our understanding is that corrections for multiple comparisons should be applied when the same data set has multiple comparisons run upon it.

We thank the reviewer for these comments. The diglossic LSBQ and the vocabulary test were only administered in order to ensure that our participants had certain levels of Arabic proficiency and use. We have now followed the reviewer’s suggestions and run two sets of regression analyses, which we report as follows, in the Results (P 24):

“4.4. Continuous experience-based factors as predictors of diglossics’ performance

Following up from studies that used measures quantifying the bidialectal experiences as continuous predictors of their participants’ performance (Poarch et al., 2019; Scaltritti et al., 2017), we ran two sets of regression analyses on the diglossics’ data using our own measures as predictors of the Flanker effect (Incongruent-Congruent), the Stroop effect (Incongruent- Congruent), and the Switch effect (Change-Stay), by adding age and sex as covariates. The first measure was diglossics’ proficiency in Modern Standard Arabic as measured by the Arabic vocabulary test. These analyses did not yield significant results for the Flanker effect (p=0.63), the Stroop effect (p=0.593), and the Switch effect (Change-stay) (p=0.63). The second measure was the diglossic LSBQ composite score. These analyses did not yield significant results for the Flanker effect (p=0.77) or the Switch effect (p=0.63); however, the model was significant for the Stroop effect (R2 = .301, F(3, 28) = 4.02, p = .017), with the LSBQ composite score demonstrating a positive correlation with the Stroop effect (p=0.04), suggesting that the higher the composite score, the higher the Stroop effect. In order to explore this further, we ran separate models to investigate the effect of the three switching domains that the LSBQ measures. The models were not significant for switching with friends (p=0.628) and for switching in social media (p=0.606), but the effect was significant when switching with family was used as the predictor (R2 = .265, F(3, 28) = 3.37, p = .033), which positively correlated with the Stroop effect (p=0.009). This may have driven the significant effect of the composite score.”

And in the discussion (p 26-27):

“An unexpected finding in our study was the positive correlation between the Stroop effect and measures of diglossic experience, namely the LSBQ composite score and the “switching with family” subscale of the same test. If anything, we would expect the opposite pattern, that is, the more experience the participants have in switching between languages, the easier it would be to process the incongruent trials, i.e. the smaller the Stroop effect would be. However, this prediction is usually drawn from findings from bilinguals not necessarily in SLCs, so it may not entirely apply to our diglossic environment and the particularities it encompasses. Taken at face value, this pattern suggests that people who switch in one context only may actually face difficulties in inhibitory control as it is measured by the Stroop task. Nevertheless, such effects that have occasionally be characterised as bilingual disadvantages (but see Luk (2022), for a criticism on this term), have been reported very rarely in the literature, and have been attributed to Type I errors (de Bruin et al., 2015). Based on this, our relatively small sample size, and the fact that such an effect was not observed for the related Flanker effect, we treat this finding with caution”.

Discussion

-We don’t necessarily disagree with the reviewer on this point. We would like to be clear that our comment on unmatched SES refers to the overall difference that we reported, not the lack of evidence for diglossic benefits. We wouldn’t know how to use the difference in SES to explain the latter actually- if anything, and assuming that there is a significant difference in SES that we did not control for, the lack of differences between our groups indicates that SES may be irrelevant for these effects. See also our response to a similar question above. We also understand that the inclusion of the claims about SES in the penultimate paragraph of the Discussion may have been confusing, so this paragraph has now been edited for clarity.

Clarity of writing

Some spaces are missing (this may have happened in the submission process or before), such as in abilingual, tolanguage, and includeexecutive on page 1.

-We apologize for these typos, and we assure that this happened during the submission process. The manuscript has been double checked for such issues, and we hope they won’t appear again in the submitted version

On page 8 under Results: Kendrick et al. (2019) are not the authors of the lme4 package, please use the appropriate reference here.

This has been fixed

Reviewer 3 Report

Please see the attached file with my comments.

Comments for author File: Comments.pdf

Author Response

This article compares diglossic speakers, or individuals who speak two varieties of the same language, to monolingual speakers on a series of executive function tasks. The results showed that diglossia did not yield an advantage on the executive function tasks, and the authors discuss these findings through the lens of the Adaptive Control Hypothesis, arguing that diglossic speakers are typically using their two versions of their language in single language contexts, with little room for overlap – hence the lack of benefit to executive function. These data are important as not much work has been done on diglossic speakers to date, and the little work that has been done has yielded conflicting findings. Overall, the manuscript was well-written. The majority of my comments are points that only require minor clarification or elaboration. I do note a few places in the Introduction section that could use further citation and explanation, particularly concerning the concepts of joint activation and a discussion about how much diglossic individuals are actively using each version of their language. I was also curious as to the effects of socioeconomic status on the results, and note that these data would be important to include and contextualize the study through, if available. This is interesting work and seems like a good fit for the special issue they are submitting to.

Introduction Comments:

Introduction, first paragraph – It seemed that there were citations missing here that may have been useful to support some of your claims. For example, you say that a bilingual’s two languages are constantly active (p. 1 line 25); this would be a good place to cite evidence from the joint activation literature, going back to some work by Viorica Marian. I similarly felt your claim on p.1, line 33-36 could use citation as well.

Thank you for this comment. We have now added some relevant references in the text

Page 2 line 52 – “summarize” instead of “summaries”

-This has been fixed

Page 5, lines 228-231 – This sentence got me thinking about the active use of one’s two languages. For the studies cited in the Introduction particularly those focused on bidialectal/biglossic cases, do they report how much participants actively use each of their two varieties of their language? In cases where one is not actively used, it would make sense to not see effects on executive functions. I would appreciate more elaboration on this point throughout the Introduction.

-This varies a lot between different studies, especially those in bilinguals (and this has been the motivation for the present study). We acknowledge that our description of our reported studies on bidialectalism and diglossia (Scaltritti, Poarch, Antoniou) was not very specific as far as this aspect is considered, so we have edited them accordingly (pages 8-10).

Method Comments:

Was the LSBQ given in standard Arabic for the diglossic speakers? How might this have influenced the way in which participants responded?

-Given that Standard Arabic is the language used for writing, the Arabic version of LSBQ was conducted in Standard Arabic. This has been clarified in the Methods and the translated version has been provided as an appendix. Regarding whether this has influenced the responses, this is hard to tell. If the reviewer is wondering about whether there were issues with how some items were translated, or whether some participants did not know any of the words, we did not receive such feedback. It would probably be worth for this test to be properly normed in Standard Arabic, something that we unfortunately did not have the resources to do.

What were participants’ scores on the Standard Arabic vocabulary test (p. 6, line 284)? A mean and standard deviation or even a median and range would be good to report.

Please see response to Reviewer 2 above

Page 6, line 292 – This should be section 3.3.1?

We have now checked the numbering of all subsections

Page 7, line 310 – I know there is an example of a neutral word in the figure, but give an example of a neutral word in the text too.

We have now added an example (“dry”) next to the description of our Stroop task

Results

Throughout the Results section, it looks like there were some degrees of freedom values missing from some of your statistical tests (for example, see p. 8, line 384 or p. 9, line 385). There are a few other instances where this occurs as well.

-Degrees of freedom have now been reported for all Reaction Time analyses; however, they are not available for generalised linear mixed effect models due to the complexity in how they are computed.

I was a bit confused by the presentation of the reaction time data for each executive function task. For example, in section 4.1.2 (p. 9, line 394) are the number of trials different for the diglossic individuals and the monolinguals because of different numbers of participants in the two groups, different numbers of trimmed trials in the two groups, or a combination of both? Perhaps make this a little clearer.

-We have now substantially changed the results section following re-analysis and recommendations from other reviewers. The number of trials are different due to the different number of participants in each group, and the different number of trimmed trials that affected the groups differently.

The placement of Figure 2 seemed odd to me, considering it showed RTs for all of the tasks, not just the Flanker. Consider putting it at the end of the Results section.

-This has been addressed.

A more general comment – did you collect socioeconomic status data on the participants? I wonder if that could be confounding the results at all. I would consider running the analyses with SES as a covariate if possible, if not I would definitely discuss as a limitation.

Rev 2 has also enquired this- please check our response above; in brief, no, unfortunately we did not have an SES measure. The probable differences in SES may be able to explain the overall difference in RTs, but not necessarily the lack of a diglossic advantage. This has been added to our Discussion as a limitation.

Discussion

Page 16, lines 615-616 – You mention that the diglossic participants were older than the monolingual participants, was this difference significant? If so, please mention it in the Method section.

-Yes, age difference was significant between our groups, and we have now reported the relevant stats. In order to ensure that there were no effects of age, also given the wide age range of our sample, our current models include Age as a covariate of no interest.

In the paragraph that begins on p. 16, line 658, you discuss how the diglossic participants in your study likely did not engage in code-switching as this is the cultural norm for them. Did you collect any switching data as part of the LSBQ (Question 22) to substantiate these claims? If so, please include any pertinent information on frequency of switching in the manuscript.

-Indeed, such data were collected as part of the LSBQ. We have now added the following information (p 14-15)

“Our monolingual young adults had an average composite score of -8.29 (SD=1.2; range: -9.66 to -4.91), meaning they could all be safely classified as monolinguals according to the LSBQ. Moreover, monolinguals scored low in all three domains of switching found in the LSBQ, including switching with family (mean: -0.93, SD: 0.53, range: -1.12 to 0.96), switching with friends (mean: -0.44, SD:0, range: -0.44 to -0.44), and switching in social media (mean: -0.45, SD: 0.9, range: -0.47 to -0.08).

To ensure that our diglossic group were indeed diglossics (knowing both the standard and the spoken variety), we adapted the LSBQ to investigate the degree of variety use, and the domains in which each variety was used (see Appendix). Given that Standard Arabic is the language used for writing and reading, the Arabic version of LSBQ was conducted in Standard Arabic. Diglossics achieved an average composite score of -0.22 (SD=2.6; range -6.62 – 4.92). According to Anderson et al., (2018), this score lies in the grey area between bilingualism and monolingualism. Crucially, results from the LSBQ revealed low scores in the three domains of switching, including with family (mean: -0.57; SD: 0.60; range -1.12 to 0.96), with friends (mean: -0.025; SD: 0.04; range (-0.44 to 1.09), and in social media (mean: 0.12, SD: 0.55; range -0.47 to 1.21). This shows that diglossics belong to an environment with limited amounts of switching (SLC); however, they significantly engage to switching more than monolinguals do in all three domains (all ps<0.05)”

Round 2

Reviewer 2 Report

The authors have put great effort into adapting their manuscript based on the previous round of reviews. Unfortunately, the authors’ explanation for comparing two completely different groups of participants has not convinced me, as there are examples (discussed in the manuscript) of groups which are much more comparable, e.g., Swabian-German bidialectals (vs. German speakers) and Italian-Venetian bidialectals (vs. Italian speakers)). My main concern - that this is a skewed and confounded comparison - therefore persists. The fact that the English monolinguals were significantly younger only makes this more apparent. Adding age as a covariate (i.e. main effect) unfortunately does not completely solve this issue (what’s more, the fact that there is a significant effect of age shows that this should not be taken for granted). A more reliable comparison might be between the older adults tested by the same authors and the younger adults from the current manuscript.

Minor comments:
“From a syntactic perspective, H is known to have a SVO word order while L has an SVO order” so they both have SVO word order? Where is the contrast in this sentence?

The reasoning behind investigating diglossia when results in bilinguals have been unreliable is still not sufficiently clear in the introduction (p. 3)

Page 14 (Section 4.4): “by adding age and sex as covariates”. It is not clear what this refers to

It is also unclear how these regression analyses were run

More information of the LSBQ is needed to interpret its (unexpected) interaction with the strength of the Stroop effect. Could this have to do with dominance (as in the Swabian-German speakers)?
It is also unclear why the predictor of switching with family has a statistical result (R-squared, p-value) and then the Stroop effect has another p-value reported.

Author Response

Dear editor and reviewer,

Thank you for forwarding the latest review of our paper. We have now addressed the new comments, and the few edits on the manuscript are highlighted in green. We hope that this version of the paper is appropriate for final acceptance. Below you can find our detailed responses:

-We thank the reviewer for this comment, but we stand behind our comparison. With respect to the Reviewer’s suggestion that we could have more comparable groups, similar to those in the Poarch and Scaltritti studies- this is virtually impossible. While in both Germany and Italy it is not difficult to find monolingual speakers of the “standard” language, in Saudi Arabia (and in general in the Arabic world), there are no monolingual speakers of Modern Standard Arabic, and monolingual speakers of the low variety at this age will have not been through formal education, and will be essentially illiterate, and potentially of a very low SES background. Comparing such a sample to our diglossic sample would cause a range of issues with regards to their compatibility. In brief, if a monolingual baseline is needed in order to assess the effects of diglossia in such an environment, this can only come from outside the Arab world.

Regarding the issue of age: we agree that this is not ideal, hence we added age as a covariate in the analysis to account for any potential effects. If the reviewer is enquiring why we did not add an interaction term, this is because we did not have a specific prediction that age would interact with performance at this age range (20s) and we did not want to produce even more complex models for no apparent reason. This was corroborated by the finding that age did not have a significant effect on the overall RTs in any of the tasks. A comparison between younger and older diglossics would indeed make sense if we were interested in the effects of age on EFs (a different research question altogether), which we would probably find; however, this would not tell us anything about the effects of diglossia- unless of course we added monolingual controls, which would ran into the other issues that the reviewer raises. Having said that, we would also be alarmed if this significant between-groups difference in age was accompanied by significant differences in the magnitude of the effects between our groups- indeed, although diglossics were overall slower than monolinguals, and this could indeed relate to them being older (this is acknowledged in the paper), the two groups did not differ in terms of the Flanker, Stroop or Switch effects. If anything, this shows that this small difference in age does not affect the speed by which incongruent trials are processed, at least at the age range of our interest. As a final note, we do recognise that it is uncommon to publish null effects, but we do not want to contribute to the purported “publication bias” in the field by dismissing this data, and we still think that our pattern of results can inform future research.

Minor comments:

“From a syntactic perspective, H is known to have a SVO word order while L has an SVO order” so they both have SVO word order? Where is the contrast in this sentence?

-Apologies, this is a typo. H has an VSO word order and L has an SVO order. This has been fixed in the manuscript

The reasoning behind investigating diglossia when results in bilinguals have been unreliable is still not sufficiently clear in the introduction (p. 3)

We believe that we are clear towards the end of the same paragraph (also in other parts of the paper and in the abstract) that Arabic diglossia here is used as a test case for the effects of using two languages/varieties in Single Language Contexts, i.e. as a situation that is appropriate to test some of the predictions of the ACH. We also maintain that diglossia in the Arab world is different than bidialectalism or indeed diglossia in Cyprus. Our argument has been that existing research on bilingualism has not adequately controlled for different conversational contexts, so here we attempt to isolate any (lack of) effects that can be attributed to SLCs, which may apply even to bilinguals in some of the previous studies

Page 14 (Section 4.4): “by adding age and sex as covariates”. It is not clear what this refers to. It is also unclear how these regression analyses were run

This refers to age and sex being added to the regression as regressors of no interest. We have reworded this section as follows in order to improve our description of the regressions, and we hope it is clearer now (page 24):

“we ran two sets of multiple regression analyses on the diglossics’ data using our own measures as predictors of the Flanker effect (Incongruent-Congruent), the Stroop effect (Incongruent- Congruent), and the Switch effect (Change-Stay). In all models, age and sex were added as regressors of no interest. The first set of analyses used as the main predictor the diglossics’ proficiency in Modern Standard Arabic as measured by the Arabic vocabulary test. These analyses did not yield significant results for the Flanker effect (R²= 0.059, F(3,28)= 0.584, p=0.63), the Stroop effect (R²= 0.065, F(3,28)= 0.644, p=0.593), and the Switch effect (Change-stay) (R²= 0.057, F(3,28)= 1.63, p=0.20). The second measure was the diglossic LSBQ composite score. These analyses did not yield significant results for the Flanker effect (R²= -0.049, F(3,28)= 0.513, p=0.677) or the Switch effect (R²= -0.042, F(3,28)= .585, p=0.63); however, the model was significant for the Stroop effect (R² = 0.301, F(3, 28) = 4.02, p = 0.017), with the LSBQ composite score demonstrating a positive correlation with the Stroop effect (Beta= 0.51, p=0.004), suggesting that the higher the composite score, the higher the Stroop effect. In order to explore this further, we ran separate models to investigate the effect of the three switching domains that the LSBQ measures on the Stroop effect. The models were not significant when switching with friends (R² = -0.041, F(3, 28) = 0.589, p=0.628) and switching in social media (R² = -0.038, F(3, 28) = 0.624, p=0.606) when used as the main predictors, but the effect was significant when switching with family was used as the predictor (R² = 0.265, F(3, 28) = 3.37, p = 0.033), which positively correlated with the Stroop effect (Beta= 0.459, p=0.009). This may have driven the significant effect of the composite score.

More information of the LSBQ is needed to interpret its (unexpected) interaction with the strength of the Stroop effect. Could this have to do with dominance (as in the Swabian-German speakers)?

The LSBQ does not provide measures of language dominance other than the measures that we have already used. We could perhaps interpret the composite score as an index of dominance, in that the higher the score the more balanced they are, and in that case the current pattern is indeed reminiscent of the one reported in Poarch et al. In any case, we have added a note acknowledging the parallels between this finding and the one reported in Poarch et al., as the Reviewer suggests (page 26):

“If the LSBQ composite score is interpreted as an index of dominance, in that higher scores suggest more balanced use of the two varieties, then this pattern is compatible with the results reported in Poarch et al. (2019), i.e. that benefits in cognitive control arise with more dominant use of one language over the other”

It is also unclear why the predictor of switching with family has a statistical result (R-squared, p-value) and then the Stroop effect has another p-value reported.

For these analyses we reported the result of the regression (which included the regressors of no interest) and also the p value of the relevant coefficient. We hope this has become clearer after our edits above, including the full stats for the regressions that did not yield a significant effect and the beta values for the significant effects.

Article Menu

Printed Edition

Is There an Effect of Diglossia on Executive Functions? An Investigation among Adult Diglossic Speakers of Arabic

Further Information

Guidelines

MDPI Initiatives

Follow MDPI