Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Please see the attached document.

Comments for author File: Comments.pdf

Author Response

 

 Review of “Functionalism and Connectionism as Foundational Theories for Usage-Based SLA:

An Explanatory Model for L2 German Case Acquisition,” Languages

Combining Functionalism and Connectivism under the umbrella of the Usage-based framework is a useful working premise for the chapter. The paper provides interesting information regarding Connectionism, Functionalism, and German case acquisition, along with discussing Processing Instruction and Concept-Based Language Instruction. However, the paper would benefit from greater use of citations (especially recent work) throughout the sections of the manuscript, clearer glossing, clearer interpretation of figures, better explanation and exemplification of key terms and concepts, and the addition of a conclusions section that summarizes the paper, explains the possible limitations of the proposed approach, and provides future directions. The writing is generally acceptable, although the manuscript would benefit from closer proofreading.

Overall, despite its current strengths, the manuscript would benefit from major revisions. Since this is a non-empirical piece, references, examples, and connections across topics are especially important. I encourage the authors in their continued pursuit of their interesting and worthwhile ideas.

Below, I first list the more major, content-related suggestions in the chronological order of the paper, followed by more legibility-oriented concerns in the next section.

Higher-level, content-related suggestions:

p.1 – line 14 – “parasitically” has pretty negative connotations. Without more context here, you might consider a different term. Later, you would benefit from explaining a bit more what you mean by this, especially for readers coming from different approaches.

Removed parasitically. Later in the paper I do use the term parasitic transfer to indicate a negative connotation towards usage of an L1 pathway, and it is explained later, but here the transfer I’m talking about is not really being used negatively so I removed it.

p.2 – lines 43 and 46. “certain communicative functions served by the first language” and “other communicative functions”, it is explained in a lengthier way in following sections, but it could be good to add at least one example so that the reader can have a better idea of different ways of creating similar functionality.

To provide examples, added “For the case of L1 English learners of German as an L2, this could be something like the use of modal verbs” and “, like thematicization and focus, which will be discussed in further detail in this paper,”

1. 2 – lines 69-72 “learners who rely on L1 pathways disregard valuable L2 grammatical components which would allow them to construct an appropriate processing model for German syntactic role disambiguation, leading to errorful interpretation and application of grammatical gender, case, and number agreement.” – Could you provide an example of this? It will help the reader.

Added: “This is a common error for L1 English learners of German as an L2, because when students encounter an accusative or dative object as the first noun, they assing it a subject role based on its syntactic position, rather than the morphological information indicating its non-subject role. For a German speaker, a sentence that begins with “Der Frau…” (“The-DAT woman…”) is indicative that the subject will come later, but learners of German often assume that woman is the subject”

p.2 – lines 74 to 78. That is a better and clearer explanation of the goals of this paper than the one stated in the abstract. Also, there is a mention of “awareness-raising tasks” that is not further developed in the paper, or at least not to the extent of Processing Instruction and Concept-Based Language Instruction.

Replaced final sentence of abstract with the one mentioned above and removed reference to “awareness-rasing tasks” in both.

1. 2 – line 87 – I think you’ll want to be careful drawing an innateness distinction between UG and functionalism. Functionalists also believe that humans are biologically disposed to acquire language. They simply differ in the nature of this capacity. For more, see Tomasello:

Tomasello, Michael (1998). ‘Introduction: A Cognitive–Functional Perspective on Language Structure’, in Michael Tomasello (ed.), The New Psychology of Language: Cognitive and Functional Approaches to Language Structure. Lawrence Erlbaum, vii-xxxi.

Tomasello, Michael (2003). ‘Introduction. Some Surprises for Psychologists’, in Michael Tomasello (ed.), The New Psychology of Language. Lawrence Erlbaum, 1-14.

Changed innate mechanisms to Language Acquisition Device

p.3 – lines 92 to 141. This is the explanation of Functionalism and how it influences SLA, and it would benefit from references, since there are no citations on this page. For additional ideas, see suggestions for citations throughout this document.

Added the following citations:

Greenhill, S. J., Wu, C. H., Hua, X., Dunn, M., Levinson, S. C., & Gray, R. D. (2017). Evolutionary dynamics of language systems. Proceedings of the National Academy of Sciences, 114(42), E8822-E8829.

Bates, E., & MacWhinney, B. (1989). Functionalism and the competition model. In E. Bates & B. MacWhinney (Eds.), The crosslinguistic study of sentence processing (pp. 73-112). New York: Cambridge University Press.

1. 3 – line 109-111: “In other cases, like English, German, and French, the marking of time is done obligatorily through morphology on the verb and can be supplemented optionally through individual lexemes.” – This is a tricky description in that it indicates that English provides temporal information obligatorily through morphology on the verb, but that it is not completely accurate. “They talk” expresses the present, but there is no morphological temporal information expressed on the verb. The sentence should be revised. Perhaps, “is often done obligatorily.”

Added “often”

p.3 – lines 119 to 120. A citation to support the greater prevalence of multilingualism would be useful.

Added citation: Grosjean, F. (2024). On bilinguals and bilingualism. Cambridge University Press.

p.3 – lines 121 to 124. I’m not quite about the appeal to a “perfect” system for either generative or functional approaches. This either will require more explanation or perhaps should be deleted.

Changed to: A Usage-based approach to understanding L1 acknowledges that the L1 system arises out of use of these multiple languages

p.3 – lines 133 to 141.You might appeal to the noticing hypothesis, since you’re alluding to some of its main contributions:

Schmidt, R. (1990) The role of consciousness in second language learning. Applied Linguistics 11, 129–58.

Added Learners need to become aware of discrepancies between the old and new systems in order to learn the target language with Schmidt citation

1. 3 line 141 – you might also note that it is successful communication that propels learners to the next acquisitional stage (see Bardovi-Harlig, 2020, p. 53). In fact, you don’t cite Bardovi-Harlig anywhere, but it would be appropriate to consider some of her seminal work on topics relevant to your chapter.

Bardovi-Harlig, K. (2017). Beyond individual form-meaning associations in L2 tense-mood-aspect research. In M. Howard & P. Leclercq (Eds.), Tense-aspect-modality in a second language: Contemporary perspectives (pp. 27–52). Benjamins. https://doi.org/10.1075/sibil.50.02bar

Bardovi-Harlig, Kathleen (2020). ‘One Functional Approach to L2 Acquisition: The Concept-Oriented Approach’, in Bill VanPatten, Gregory Keating and Stefanie Wulff (eds), Theories in Second Language Acquisition, (3rd ed.). Routledge, 40-62.

Added relevant quote from Bardovi-Harlig, Kathleen (2020) and post-citation description of its relevance to this part of the text

p.4 – lines 144 to 170. Connectionism receives much less room/attention here than Functionalism. This part should be at least as lengthy and detailed as section 1, which will allow you to get into more detail re: cue competition, etc.

added : Within the Competition Model framework (CM) (Bates & MacWhinney, 1989), cues are seen as information that helps speakers determine the relationship between forms and meaning. These cues compete for activation and their strengths or cue-weights are always intertwined as part of their relationship to other cue. The weight or strength of a particular cue is developed in two ways. First is by its validity, which consists of its availability in the input and its reliability for predicting the appropriate form-meaning connection. And second, by its relative validity to other competing cues. Much of the early work on the CM focused on child-language learning, although from its concep-tion, it was always boldly intended as a way to capture the entirety of language learn-ing, use, and attrition in both comprehension and production.

and: The Unified Competition Model (UCM) (MacWhinney, 2005) specifically shows how both L1 and L2 acquisition and learning result from the same fundamental principles of cue competition and further takes into account the impacts of transfer between linguistic codes. The UCM covers cue competition in the different “arenas” in which competition can play out, including phonology, the lexicon, morpho-syntax and conceptual knowledge. As MacWhinney (2005) argues, “forms compete to express underlying intentions or functions,” (71) in production, but also for interpreting intentions and functions in comprehension. The UCM also takes into account the effects of chunking, or the size of cues, grammatical forms, and lexical bundles, as well as the necessary storage of these cues for future activation. And finally, the UCM also implicates the importance of code activation, i.e. which linguistic units of the entire linguistic system should be activated based on the language being spoken. In this model, a particular code does not distinguish necessarily between two separate linguistic subsystems, as more than one language may utilize the same parts of the linguistic subsystem, which allows us to understand the mechanism behind transfer effects, as well as the need in L2 learning to build differentiations between linguistic codes where necessary.

 

p.4 – lines 158 and 159. It would be nice to briefly explain and/or exemplify “U-shaped learning” and “interlanguage transfer” at these initial uses so that it is easier for the reader to understand their relevance in this section.

Changed to: U-shaped learning, i.e. the pattern of rote learning of irregular forms replaced by overgeneralization of regular patterns onto irregular lexems and eventual return to correct irregular forms (Ellis & Wulff, 2020) and interlanguage transfer, i.e. the adoption of L1 cues and cue-weights in L2 language processes as well as the opposite, where learned L2 cues and cue-weights may affect L1 processes (MacWhinney, 2018).

1. 4 lines 156-9: “For SLA specifically, competition within connectionist models can explain a wide arrange of L2 acquisitional and use patterns (Shirai, 2018), including U-shaped learning (Ellis & Wulff, 2020) and interlanguage transfer (MacWhinney, 2018).” I think the phenomenon of restructuring would also be useful to appeal to here. For a recent example, see:

Kanwit, M. (2024). Restructuring in L2 Spanish future-time variation: The evolving roles of contextual sensitivity and individual variability in morphological development over time. Journal of the European Second Language Association, 8(1), 97-113. https://doi.org/10.22599/jesla.104

added: ), and restructuring, i.e., changes to the functions assigned to already learned forms and their integrations of new forms and functions into the linguistic system (Ellis & Wulf, 2020; Kanwit, 2024).

1. 4 lines 168-70: “Under a broad, Usage-based framework, Connectionism provides the viable hardware on which observed psycholinguistic phenomena can operate.” – I think you make a good point, but I think it would help the reader to be more concrete with your summary sentence here in terms of specifics and examples. Tell or remind the reader -- what is the hardware? Give an example for a particular psycholinguistic phenomenon.

Changed to: Under a broad, Usage-based framework, Connectionism provides emphasizes the via-ble hardware, the biological, interconnected system of the brain, on which observed psycholinguistic phenomena like U-shaped learning, transfer, and restructuringcan operate.

p.4 – lines 173 to 174. This sentence is difficult to interpret. Please clarify.

Changed to: When anew communicative need arises, there are a number of linguistic ways for speakers to meet it.

1. 4 line 176 – “suprasegmental phonetic means”-> I think most would say these are “suprasegmental phonological means”

changed to phonological

1. 4 lines 183-5: “predominance of first-noun-as-subject in English will almost always be used in ambiguous cases under these linguistic constraints.” You might benefit from referring the reader to other factors that play a role in disambiguation and in determining or overriding the First Noun Principle. See chapter 6 by VanPatten (2020), pp. 111-115 in:

VanPatten, Bill, Gregory Keating, and Stefanie Wulff (eds) (2020). Theories in Second Language Acquisition (3rd ed.). Routledge.

Added:  (see VanPatten (2020) for other factors can also play a role in disambiguation and either accepting or overriding this First Noun Priniciple)

pp.4-5 lines 190-1: “German speakers do not rely on syntactic position to disambiguate the subject from other potential candidates, but rather indicative morphology for subjects” – I think you mean “nominative morphology for subjects”, which would be clearer for the reader.

Changed to nominative

1. 5 lines 199-206: You would benefit from following the glossing system of Haspelmath and Sims. You shouldn’t put dashes in the interlinear gloss (second line) when there is no gloss on the source language line (first line). In those cases, you should use a period in the interlinear gloss (e.g., the.NOM). Check the style sheet. Otherwise, to be more Leipzig-like you might do the.MASC and so forth. See Haspelmath and Sims (2010, end of Ch. 2) Understanding morphology for more on glossing guidelines. Also, in examples 2 and 3, particular items in your source and interlinear lines don’t line up, so be sure to place tabs and spaces as appropriate.

Changed to follow the.NOM style

1. 5 lines 215-7 Good point.

Thanks ?

p.6 – line 250 and 251. Provide citations for “the compelling evidence”.

Added references to Jackson 2007;2008, VanPatten and Borst 2012, Rankin 2014

p.6 – lines 255-8. It would help the reader to gloss these examples in the same way as the ones on page 5, so they can see which elements are located in which part of the clause.

 added them in style of examples sentences 1,2,3. Changed to 4, 5, 6, examples with intergloss style .NOM

p.6 – lines 263 to 267. This might be an appropriate place to engage with the concepts of scaffolding and the Zone of Proximal Development. See also the chapter by Lantolf and colleagues:

Lantolf, James, Matthew Poehner, and Steven Thorne (2020). ‘Sociocultural Theory and L2 Development’, in Bill VanPatten, Gregory Keating and Stefanie Wulff (eds), Theories in Second Language Acquisition (3rd ed.). Routledge, 108-24.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.

I didn’t think this was the right place to fit this in, as I talk about this further below in relation to C-BLI and this part is not really about instruction.

p.6 – lines 283 to 289. Please fix the spacing of the glosses so they can help the reader.

Fixed spacing. Changed 4,5,6 to 7,8,9. Changed notation to be .NOM style. Left 9, as I think a word-for-word intergloss is unnecessary

Line 301-2: For more on the role of morphology in aiding processing, see Haspelmath (2014) and Mondorf (2014), and specifically for grammatical gender, see Zubin and Koepcke (1981). Each would be helpful resources for what you note here. And, specifically, Haspelmath is a major contributor to functional morphology and doesn’t appear in the manuscript:

Haspelmath, M. (2014). On system pressure competing with economic motivation. In B. MacWhinney, A. L. Malchukov, & E. A. Moravcsik (Eds.), Competing motivations in grammar and usage (pp. 197-208). Oxford University Press.

Mondorf, B. (2014). Apparently competing motivations in morphosyntactic variation. In B. MacWhinney, A. L. Malchukov, & E. A. Moravcsik (Eds.), Competing motivations in grammar and usage (pp. 209-228). Oxford University Press.

Zubin, D. A., & Koepcke, K. (1981). Gender: A less than arbitrary category. Chicago Linguistics Society, 17, 439-449.

Added: For more on the role of morphology in aiding processing, see Haspelmath (2014) and Mondorf (2014), and specifically for grammatical gender, see Zubin and Koepcke (1981). And in the references

p.6 – Figure 1 / p.7 – Figure 2. I think one of the points you’re trying to make is that der in Figure 2 could be referring to many more types of referents (in terms of word class, number, gender) than das in Figure 1. However, you don’t ever actually say this. If this is one of your main takeaway messages, it should be stated in the body text.

Added: In Figure 1, it is clear that the determiner das is only associated with the neuter grammatical gender, only with singular nouns, and only with nominative and accusative cases. In contrast, Figure 2 shows that the determiner der can be associated with masculine and feminine and plural nouns depending in the case, making ease of processing and syntactic role disambiguation nigh impossible without knowledge of the noun gender or plural forms. This increase in processing effort is only worth the payout of easier prediction if the case-number-gender paradigm is easily accessible, especially for oral production and comprehension.

1. 8 – lines 338-341 – “they are either based off of or align with implicit, theoretical linguistic notions of innatism which lack explanatory power of underlying, biologically plausible mechanisms for the learning that occurs in these studies.” You critique these approaches in a strong fashion, but you don’t provide much detail here. Perhaps you can provide a quote or citation for connections to innatism with each of the three? Otherwise, it doesn’t quit seem fare to make the critical claim without examples or references.

Made it less harsh. I don’t think it’s necessary to go into all that here: And while Chapelle (2009) categorizes each of these three theories as “psycholinguistic”, they are ambivalent in their direct connection to underlying, biologically plausible mechanisms for the learning that occurs in these studies.

 

Line 354-5 – van Compernolle is one of the leaders in C-BLI and isn’t cited in the manuscript (unless anonymized). Similarly, Eskildsen has advocated for integrating usage-based approaches and a sociocultural orientation, and Kanwit notes the benefits of combining variationist sociolinguistic and concept-oriented approaches. Consider, for example, works such as:

Eskildsen, S. W. (2015). What counts as a developmental sequence? Exemplar-based L2 learning of English questions. Language Learning, 65, 33-62.

Kanwit, M. (2017). What we gain by combining variationist and concept-oriented approaches: The case of acquiring Spanish future-time expression. Language Learning, 67(2), 461-498.

van Compernolle, R. A. (2019). Constructing a second language sociolinguistic repertoire: A sociocultural usage-based perspective. Applied Linguistics, 40(6), 871-893.

Added citations for C-BLI, and the following for their integration in Usage-based theories: And while some recent studies have argued for the integration of C-BLI and Sociocultural approaches to SLA with Usage-based theories (e.g., Eskildsen, 2015; Kanwit, 2017; van Compernolle, 2019), further work is needed to articulate the ways in which these theories can be fully aligned

Lines 359-360: “Functionality is foundational to Usage-based approaches, as it provides the motivation for linguistic forms and their propagation to fulfill communicative needs..” yes, see Bardovi-Harlig (2018, p. 190) for similar, compatible claims:

Bardovi-Harlig, K. (2018). Concept-oriented analysis: A reflection on one approach to studying interlanguage development. In A. Edmonds & A. Gudmestad (Eds.), Critical reflections on data in second language acquisition (pp. 171-195). Amsterdam: John Benjamins.

Added citation

Lines 364-5: “true for both statistical learning of specific L2 forms, as can be modelled to exhibit similar patterns to L2 learning outcomes, like U-shaped learning” – An example would be useful here.

Already provide an example of U-shaped learning earlier and since this is the conclusion, I’m not sure its necessary. I have not changed this.

Lines 372 – This is the first time the phrase “cue validity” appears in the manuscript, so it’s odd to have it so late as an example related to a hypothesis that your proposal can offer. You should say more about these examples you’ve listed in this particular sentence.

I added information earlier about cue validity.

1. 9 – Your manuscript ends rather abruptly. It would be beneficial to provide one last section in which you summarize the chapter, list limitations of this combined framework and provide future directions for ensuing studies.

I separated the last paragraph into two. I disagree that the last paragraph ends abruptly and think it works well with this type of essay-style paper. I’m not sure that “limitations” is necessary, as I explain what is testable and provide an example. I decided to keep the last paragraph, now split into two, basically as is.

Further suggestions for greater legibility:

p.1 – lines 6 to 9 are the exact same as lines 31 to 34 of the same page, it would be beneficial to reword one of the two.

Left as is because 6-9 is in the abstract…

p.1 – line 25 – “show” -> “show that” to aid legibility

already changed based on comments abos

p.1 – line 39 – “are neural basis”- > “is a neural basis”

changed to is a neural basis

p.2 – line 47 – add comma after “force”

comma added

p.2 – line 53 – add comma after “Thus”

comma added

p.2 – lines 54 and 55. “processing” is repeated twice in the sentence, and the repetitions are very close to each other. It would be better to reword or use a synonym.

p.2 – lines 56 to 59 are the exact same as lines 19 to 22 of the same page, it would be beneficial to reword one of the two.

Again, from the abstract vs body of paper. Keeping the same.

1. 2 – line 65 – “its function” – I think you mean the case-marking system’s function, but this isn’t entirely clear, so please reword

changed to the case-marking system’s function

1. 2 – line 66 – “syntactic role” -> “syntactic roles”

changed to roles

p.2 – line 90. “Because of our species capacity…” -> “Because of our species’ capacity…”

apostrophe added

1. 2 lines 94-98 – “In the continued refinement and specification of speakers’ needs over time, we can find the dynamic and historical processes that have differentiated various linguistic forms within languages to serve these purposes, or functions. As a new communicative need arises, the linguistic systems that enable communication are adapted by speakers to fill the gap and the new forms take their place not in isolation from other forms but rather gaining meaning in contrast to and connection with previously established forms.” – This is a rather passive way to describe this process, taking some of the agency away from learners, I would reword to active voice throughout.
2. 3 line 117 – Unclear referent for “they”

changed 2^nd sentence to be active As a new communicative need arises, speakers adapt the linguistic systems that enable communication to fill the gap and these

1. 5 238-9: “For this reason, it is vital that L1 English L2 German learners are exposed early and often to non-SVO word order and are positioned to be forced to processes them for appropriate syntactic role.” – “to processes”-> “to process”, “Them” has an unclear referent, and you should add “the” before “appropriate syntactic role”

changed to For this reason, it is vital that L1 English L2 German learners are exposed early and often to non-SVO word order and are positioned to be forced to process non-SVO sentences for the appropriate syntactic role.

Line 247: “to at least processes”-> “to at least process”

Changed to to at least process

1. 8 line 334 – “learners processing” -> “learners’ processing”

apostrophe added

335. 8 – line 335. “The theoretical underpinnings of PI are rooted in Processing Instruction”-> “…rooted in Input Processing”

Changed to Input Processing

Lines 365-6: “as well as provide”- > “as well as to provide”

Added: to

Reviewer 2 Report

Comments and Suggestions for Authors

General Comments

The paper serves as an introduction to Functionalism and Connectionism and uses case-marking in German as a case study, explaining how learning challenges associated with the form can be modeled and explained within the context of these usage-based approaches. Finally, the paper discusses pedagogical approaches to this form with an eye towards how results in the Processing Instruction and Concept-Based Language Instruction literatures can be explained under a usage-based account.

In general, I found this to be an engaging and insightful work that makes some important points and draws some important connections between different literatures. At the same time, I feel like some of the discussion could be meaningfully expanded so that these connections are clearer and more impactful.

In the comments that follow, I’ve listed several major comments, that I feel need to be addressed before publication. I have also listed more minor comments that should help to clarify some points that arose in my reading. Finally, I note a few typos or stylistic issues. These are listed with reference to the section or line numbers (as numbered in the manuscript PDF).

 

Major Comments

1. Section 3. In this section, the focus is on how case operates using an example from the dative case. However, this example, and the ensuing discussion completely ignores the accusative case. This is extremely important to the phenomenon being discussed here for several reasons. First, because feminine, neuter, and plural nouns are the same in the nominative and accusative cases, case does not actually indicate grammatical roles in sentences like “Die Katze sieht das Kind.” Here, grammatical role is indicated by word order (or other cues, such as animacy, prosody, etc.) because case is entirely uninformative. Relatedly, while the discussion in 215-225 lists multiple reasons why SVO items pose a challenge for L2 learners (“The learner, having encountered no issues utilizing the L1 pathway for subject role assignment, will be more apt to process sentences this way in the future…”), it is important to point out that the L1 pathway is strengthened not just by those instances where it is possible to use word order to interpret the sentence correctly, but by those instances where it is necessary to do so.
2. 251-261. In this section, a relative lack of non-SVO constructions is discussed (and this is brought in relation to lack of knowledge of case marking). First, I’m not convinced that a lack of XVS constructions is the same as a lack of OVS sentences. That is, a lack of knowledge of case marking is certainly responsible for the lack of OVS sentences. But I’m not sure it has anything to do with the lack of XVS sentences, where a time or location fills the ‘X’. Can this relationship be explained? Further, it is stated that “Even in cases where learners are able to process German sentences in XVS order, they often fail to reproduce this syntax, revealing a continued use of L1 production routes”. I’m not convinced that the failure to reproduce S-V inversion can be attributed (solely) to “a continued use of L1 production routes” given that research on acquisition orders and developmental stages has shown that L1 learners of German and L2 learners from a variety of language backgrounds also produce such structures (Clahsen, 1984).
3. Section 4. As I understand it, this section aims to (a) describe findings from Processing Instruction and C-BLI, (b) note why the theoretical underpinnings “leave something to be desired”, and (c) explain how usage-based models can fill the gap. With regard to (a), the discussion of actual findings is rather light—they are simply summarized in one sentence. With regards to (b) and (c), I also think there is some discussion missing.
   
   As my own work is in PI, not C-BLI, I’ll highlight what I see as the shortcomings of this discussion from that perspective, though I suspect someone with a background in C-BLI might have similar comments with reference to that paragraph: First, I’d note that in the discussion of PI, it is simply stated that the theories underpinning PI “lack explanatory power of underlying, biologically plausible mechanisms for the learning that occurs”. But here, there is nothing offered in the way of explaining why the accounts given in this work are insufficient (I might, for example, look at VanPatten & Rothman (2014) for a detailed account of the mechanism said to be at work). Second, the discussion glosses over the fact that work on PI has occasionally been discussed as theory-independent (despite VanPatten’s clear orientation towards UG frameworks), and some works have made explicit attempts to combine insights from the PI literature with usage-based psycholinguistic approaches (e.g., Henry, 2023). I’d highlight two quotes to illustrate the point: VanPatten (2015) says: “To be sure, I take a formal approach to the nature of the underlying system, but as I state in various publications, one need not take such a perspective to understand how PI might help language grow in the mind/brain (see, for example, VanPatten 2009, 2014).” With reference to VanPatten’s lexical processing principle, Dracos & Henry state: “Although the underlying theoretical assumptions are different, models of learned attention and cue blocking make similar predictions given that salient lexical and discourse cues typically overshadow verbal morphology in natural learning contexts (see, e.g., Ellis and Sagarra 2010a, 2010b)” (Dracos & Henry, 2021, p. 2).
   
   In general, I felt like this section didn’t “pay off” the way it could have. The ideas are here, and I do appreciate the explicit attempt to bridge the gaps between the PI / C-BLI worlds and functionalism / connectionism, but I think the conversation needs to be expanded in these ways in order to really make an impact.

 

Smaller Comments

1. 53, 59, 165 & others(?). At several point, the text refers to “noticing”, i.e,. “notice and adopt the L2 processes”, or “noticing form-function relationships”. The term “noticing” is potentially loaded given its potential association with the noticing hypothesis, which implies a particular type of explicit / conscious process, which may not be implicated here. Please consider revising or clarifying as necessary to account for this.
2. 59 (and 23, in the abstract). “…noun form dictates subject and objects in utterances.” Given the issue I point to above—that case does not always indicate subject / object relationships—please consider revising these statements.
3. 112-113. I’m not sure I understand this statement: “For L2 learners, this motivational basis for communication is no longer rooted in their development of their L1(s)”. (Note also it looks like there is a double space before “L1(s)”).
4. 135-135. “The wholesale adoption of existing L1 processes could be viewed as the initial state of L2 language learning.” This is, of course, one possibility, but it is not the only possibility (see, e.g., VanPatten’s discussion of the “universality” of the First Noun Principle in VanPatten (2015)). Please clarify whether this is the position you are taking or whether you suggest it as a possibility. (And also consider noting the possibility that the FNP is universal at 215.)
5. It is noted that “learners deepen the incorrect entrenched pathway…”. I would not characterize SVO sentences—or even using word order to process grammatical relations—as “incorrect.” Sub-optimal, maybe (to borrow a term from VanPatten). But given the issue I raise in major comment (1), this cannot be “incorrect” in all instances.
6. “most L2 learners … do not exist in a rich target language environment.” What is meant by “rich” in this context? I’d suppose that it’s the lack of pragmatic contexts in which OVS (or XVS) structures would typically occur?
7. “The theoretical underpinnings of PI are rooted in Processing Instruction…”. I believe what is meant is that the underpinnings of PI are rooted in VanPatten’s Input Processing model?
8. At some point in this discussion, I would have expected that Kempe & MacWhinney (1998, 1999) might have been cited in this paper given their obvious relevance to the questions at hand. I would assume that the authors are aware of these works, but if not, I’d suggest that they should be, and consider working some of their insights into the body of the text as they see fit.

 

Typos, style, formatting, and other small suggestions

1. Consider replacing the word “replicate” with “represent”.
2. 120-121. “… and as such, ____ should be the basis for…”. Insert “this”.
3. 227-228. “definitive article” à “definite article”.

 

References

Clahsen, H. (1984). The acquisition of German word order: A test case for cognitive approaches to L2 development. Second Languages: A Cross-Linguistic Perspective, 219–242.

Dracos, M. J., & Henry, N. (2021). The role of task-essential training and working memory in offline and online morphological processing. Languages, 6(1), 24.

Henry, N. (2023). The use of blocking and inhibition training in processing instruction. International Review of Applied Linguistics in Language Teaching, 61(3), 791–817.

Kempe, V., & MacWhinney, B. (1998). The acquisition of case marking by adult learners of Russian and German. Studies in Second Language Acquisition, 20(04), 543–587. https://doi.org/10.1017/S0272263198004045

Kempe, V., & MacWhinney, B. (1999). Processing of morphological and semantic cues in Russian and German. Language and Cognitive Processes, 14(2), 129–171.

VanPatten, B. (2015). Foundations of processing instruction. International Review of Applied Linguistics in Language Teaching, 53(2), 91–109. https://doi.org/10.1515/iral-2015-0005

VanPatten, B., & Rothman, J. (2014). Against “Rules.” In A. G. Benati, C. Laval, & M. J. Arche (Eds.), The grammar dimension in instructed second language learning (pp. 23–39). Bloomsbury.

 

Comments for author File: Comments.pdf

Author Response

 

 General Comments

The paper serves as an introduction to Functionalism and Connectionism and uses case-marking in German as a case study, explaining how learning challenges associated with the form can be modeled and explained within the context of these usage-based approaches. Finally, the paper discusses pedagogical approaches to this form with an eye towards how results in the Processing Instruction and Concept-Based Language Instruction literatures can be explained under a usage-based account.

In general, I found this to be an engaging and insightful work that makes some important points and draws some important connections between different literatures. At the same time, I feel like some of the discussion could be meaningfully expanded so that these connections are clearer and more impactful.

In the comments that follow, I’ve listed several major comments, that I feel need to be addressed before publication. I have also listed more minor comments that should help to clarify some points that arose in my reading. Finally, I note a few typos or stylistic issues. These are listed with reference to the section or line numbers (as numbered in the manuscript PDF).

Major Comments

1. Section 3. In this section, the focus is on how case operates using an example from the dative case. However, this example, and the ensuing discussion completely ignores the accusative case. This is extremely important to the phenomenon being discussed here for several reasons. First, because feminine, neuter, and plural nouns are the same in the nominative and accusative cases, case does not actually indicate grammatical roles in sentences like “Die Katze sieht das Kind.” Here, grammatical role is indicated by word order (or other cues, such as animacy, prosody, etc.) because case is entirely uninformative. Relatedly, while the discussion in 215-225 lists multiple reasons why SVO items pose a challenge for L2 learners (“The learner, having encountered no issues utilizing the L1 pathway for subject role assignment, will be more apt to process sentences this way in the future…”), it is important to point out that the L1 pathway is strengthened not just by those instances where it is possible to use word order to interpret the sentence correctly, but by those instances where it is necessary to do so.

Added: This first-noun-as-subject principle is not necessarily incorrect as a syntactic-role disambiguation strategy in German either, and is sometimes necessary in cases where case marking morphology is ambiguous and context does not provide any further cues. However, this produces a serious problem for the learner’s continued development with German. First, correct interpretation of the subject via the incorrect cue strengthens an improper L2 processing route…

 

2. 251-261. In this section, a relative lack of non-SVO constructions is discussed (and this is brought in relation to lack of knowledge of case marking). First, I’m not convinced that a lack of XVS constructions is the same as a lack of OVS sentences. That is, a lack of knowledge of case marking is certainly responsible for the lack of OVS sentences. But I’m not sure it has anything to do with the lack of XVS sentences, where a time or location fills the ‘X’. Can this relationship be explained? Further, it is stated that “Even in cases where learners are able to process German sentences in XVS order, they often fail to reproduce this syntax, revealing a continued use of L1 production routes”. I’m not convinced that the failure to reproduce S-V inversion can be attributed (solely) to “a continued use of L1 production routes” given that research on acquisition orders and developmental stages has shown that L1 learners of German and L2 learners from a variety of language backgrounds also produce such structures (Clahsen, 1984).

Added/changed: While only OVS constructions indicate case on the object, other XVS constructions, like time adverbials and prepositional phrases, force learners to into positions where the subject appears after the verb, which contrasts with English sentences processing where the subject always appears before the verb. Even when learners are exposed to XVS and OVS structures, the validity of SVO word order in German makes the uptake of case marking difficult, in addition to the fact that English speakers are not used to processing articles for information beyond definiteness and number (see Hopp, 2016), although other non-L1 specific processes may also be at play that prefer certain syntactic acquisitional orders (see Clahsen, 1984).

 

3. Section 4. As I understand it, this section aims to (a) describe findings from Processing Instruction and C-BLI, (b) note why the theoretical underpinnings “leave something to be desired”, and (c) explain how usage-based models can fill the gap. With regard to (a), the discussion of actual findings is rather light—they are simply summarized in one sentence. With regards to (b) and (c), I also think there is some discussion missing. As my own work is in PI, not C-BLI, I’ll highlight what I see as the shortcomings of this discussion from that perspective, though I suspect someone with a background in C-BLI might have similar comments with reference to that paragraph: First, I’d note that in the discussion of PI, it is simply stated that the theories underpinning PI “lack explanatory power of underlying, biologically plausible mechanisms for the learning that occurs”. But here, there is nothing offered in the way of explaining why the accounts given in this work are insufficient (I might, for example, look at VanPatten & Rothman (2014) for a detailed account of the mechanism said to be at work). Second, the discussion glosses over the fact that work on PI has occasionally been discussed as theory-independent (despite VanPatten’s clear orientation towards UG frameworks), and some works have made explicit attempts to combine insights from the PI literature with usage-based psycholinguistic approaches (e.g., Henry, 2023). I’d highlight two quotes to illustrate the point: VanPatten (2015) says: “To be sure, I take a formal approach to the nature of the underlying system, but as I state in various publications, one need not take such a perspective to understand how PI might help language grow in the mind/brain (see, for example, VanPatten 2009, 2014).” With reference to VanPatten’s lexical processing principle, Dracos & Henry state: “Although the underlying theoretical assumptions are different, models of learned attention and cue blocking make similar predictions given that salient lexical and discourse cues typically overshadow verbal morphology in natural learning contexts (see, e.g., Ellis and Sagarra 2010a, 2010b)” (Dracos & Henry, 2021, p. 2). In general, I felt like this section didn’t “pay off” the way it could have. The ideas are here, and I do appreciate the explicit attempt to bridge the gaps between the PI / C-BLI worlds and functionalism / connectionism, but I think the conversation needs to be expanded in these ways in order to really make an impact.

Added: citation for Henry, N. (2023). The additive use of prosody and morphosyntax in L2 German. Studies in Second Language Acquisition, 45(2), 348-369.

Added: According to VanPatten (2017), “PI is actually a type of focus on form or better yet, a pedagogical intervention. As such it is not a method with an underlying approach but instead an intervention that can be used by any communicative approach that seeks a supplemental or periodic focus on the formal features of language,” (p. 166). From this perspective, PI should be applicable under any number of language theories. However, in the same paper, VanPatten (2017) declares that “Under PI, acquisition consists of three necessary ingredients: input; Universal Grammar (UG) and internal mental architecture; and processing mechanisms that mediate between input and UG/internal architecture,” (p 167). Here, there is a sharp contrast to the theory-neutral stance of its applicability and a declaration of the need for UG to explain the learning mechanism. The major challenge here is that the definition of UG has changed significantly over the past decades (see Trettenbrein (2015) for an overview) , and our understanding of human cognition has developed to handle the difficulties that UG hoped to solve in more empirical ways.

Changed: Despite the UG foundations of VanPatten’s PI, it has been shown to be effective. Other researchers take up the more theory-neutral side of PI and have investigated it along lines that explore PIs connection to other theories of langauge, such as Henry (2025), who describes PI as having two “core tenets […]

Added: distances itself to some degree fromleaves something to be desired regarding underly-ing, biologically plausible mechanisms for L2 learning, and does not necessarily disre-gard, but does not attend significantly enough to the properties we see in implicit and statistical learning that are still required for L2 development. The procedures involved in C-BLI are highly effective at helping students understand form-meaning connec-tions, notice forms, apply relevant higher-order thinking, and ultimately apply this in-formation, but it does not help us understand the route between this explicit infor-mation and automatization and/or proceduralization of this knowledge into implicit knowledge. .

 

Smaller Comments

1. 53, 59, 165 & others(?). At several point, the text refers to “noticing”, i.e,. “notice and adopt the L2 processes”, or “noticing form-function relationships”. The term “noticing” is potentially loaded given its potential association with the noticing hypothesis, which implies a particular type of explicit / conscious process, which may not be implicated here. Please consider revising or clarifying as necessary to account for this.

A reference to Schmidt was added.

2. 59 (and 23, in the abstract). “…noun form dictates subject and objects in utterances.” Given the issue I point to above—that case does not always indicate subject / object relationships—please consider revising these statements.

Added: frequently dictates

3. 112-113. I’m not sure I understand this statement: “For L2 learners, this motivational basis for communication is no longer rooted in their development of their L1(s)”. (Note also it looks like there is a double space before “L1(s)”).

Added: their L1(s); they already have a well-enough established linguistic system for communication. Rather, ….

And removed extra space

4. 135-135. “The wholesale adoption of existing L1 processes could be viewed as the initial state of L2 language learning.” This is, of course, one possibility, but it is not the only possibility (see, e.g., VanPatten’s discussion of the “universality” of the First Noun Principle in VanPatten (2015)). Please clarify whether this is the position you are taking or whether you suggest it as a possibility. (And also consider noting the possibility that the FNP is universal at 215.)

Added : , recognizing other generalized linguistic, and even non-linguistic processing mechanisms may be at play

I hesitate to, and think do not want to, recognize “universals” in this way in this paper, as they are emergent properties, even if they end up as similar properties in most normally-developing individuals. The theoretical stance behind the term “universal” is loaded in a way contrary to the arguments in this paper.

5. 253. It is noted that “learners deepen the incorrect entrenched pathway…”. I would not characterize SVO sentences—or even using word order to process grammatical relations—as “incorrect.” Sub-optimal, maybe (to borrow a term from VanPatten). But given the issue I raise in major comment (1), this cannot be “incorrect” in all instances.

Changed “incorrect” to sub-optimal

6. 262. “most L2 learners … do not exist in a rich target language environment.” What is meant by “rich” in this context? I’d suppose that it’s the lack of pragmatic contexts in which OVS (or XVS) structures would typically occur?

Changed to: In addition, most L2 learners are exposed to simplistic sentences and unlike L1 learners, do not frequently participate in rich target language environments where this word-order variability is displayed in pragmatic, contextualized ways.

7. 336. “The theoretical underpinnings of PI are rooted in Processing Instruction…”. I believe what is meant is that the underpinnings of PI are rooted in VanPatten’s Input Processing model?

Changed to Input Processing

8. At some point in this discussion, I would have expected that Kempe & MacWhinney (1998, 1999) might have been cited in this paper given their obvious relevance to the questions at hand. I would assume that the authors are aware of these works, but if not, I’d suggest that they should be, and consider working some of their insights into the body of the text as they see fit.

I’m definitely aware of the work, but as it’s mostly a cross-linguistic comparison of L2 German vs L2 Russian, I don’t think it fit’s within the scope of this paper, although many relevant ideas discussed in those papers are already presented via other citations.

 

Typos, style, formatting, and other small suggestions

1. Consider replacing the word “replicate” with “represent”.

Changed to: represent

2. 120-121. “… and as such, ____ should be the basis for…”. Insert “this”.

Added: this

3. 227-228. “definitive article” → “definite article”.

Changed to: definite

 

References

Clahsen, H. (1984). The acquisition of German word order: A test case for cognitive approaches to L2 development. Second Languages: A Cross-Linguistic Perspective, 219–242.

Dracos, M. J., & Henry, N. (2021). The role of task-essential training and working memory in offline and online morphological processing. Languages, 6(1), 24.

Henry, N. (2023). The use of blocking and inhibition training in processing instruction. International Review of Applied Linguistics in Language Teaching, 61(3), 791–817.

Kempe, V., & MacWhinney, B. (1998). The acquisition of case marking by adult learners of Russian and German. Studies in Second Language Acquisition, 20(04), 543–587. https://doi.org/10.1017/S0272263198004045

Kempe, V., & MacWhinney, B. (1999). Processing of morphological and semantic cues in Russian and German. Language and Cognitive Processes, 14(2), 129–171.

VanPatten, B. (2015). Foundations of processing instruction. International Review of Applied Linguistics in Language Teaching, 53(2), 91–109. https://doi.org/10.1515/iral-2015-0005

VanPatten, B., & Rothman, J. (2014). Against “Rules.” In A. G. Benati, C. Laval, & M. J. Arche (Eds.), The grammar dimension in instructed second language learning (pp. 23–39). Bloomsbury.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

See attached file.

Comments for author File: Comments.pdf

Author Response

See attached file

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The revisions to this piece have been thoughtfully done and have addressed the core of all of the issues I raised. There are two lingering issues I'd ask the reviewer to address, but otherwise, I think the paper makes a substantive contribution to the literature. As I stated in my original review, I found this to be an engaging and insightful work that makes some important points and draws some important connections between different literatures. With these modifications, I can recommend publication. 

1.

Original Comment

Section 3. In this section, the focus is on how case operates using an example from the dative case. However, this example, and the ensuing discussion completely ignores the accusative case. This is extremely important to the phenomenon being discussed here for several reasons. First, because feminine, neuter, and plural nouns are the same in the nominative and accusative cases, case does not actually indicate grammatical roles in sentences like “Die Katze sieht das Kind.” Here, grammatical role is indicated by word order (or other cues, such as animacy, prosody, etc.) because case is entirely uninformative. Relatedly, while the discussion in 215-225 lists multiple reasons why SVO items pose a challenge for L2 learners (“The learner, having encountered no issues utilizing the L1 pathway for subject role assignment, will be more apt to process sentences this way in the future…”), it is important to point out that the L1 pathway is strengthened not just by those instances where it is possible to use word order to interpret the sentence correctly, but by those instances where it is necessary to do so.

Author Response

Added: This first-noun-as-subject principle is not necessarily incorrect as a syntactic-role disambiguation strategy in German either, and is sometimes necessary in cases where case marking morphology is ambiguous and context does not provide any further cues. However, this produces a serious problem for the learner’s continued development with German. First, correct interpretation of the subject via the incorrect cue strengthens an improper L2 processing route…

Reviewer Response

 The addition here is appreciated, though I would suggest a few additional modifications: (a) providing an example of a sentence with a feminine / neuter in the accusative (parenthetically would be fine), and (b) changing the text in the next sentence. My suggestion would be: “Correct interpretation… via the incorrect cue strengthens an improper L2 processing route.” à “Correct interpretation via word order strengthens this route, especially given that the vast majority of transitive sentences in German can be interpreted in this way.”  My overall concern here is that using word order isn’t “incorrect” or “improper” in any real sense, though it’s not the most reliable cue when case is available. More importantly, I think the use of the examples here and the ensuing discussion underplay how often the word order cue is necessary (Kempe & MacWhinney 1998 estimate that 44% of German transitive sentences do not contain a masculine noun).

 

2. 

Original Comment

Section 4. As I understand it, this section aims to (a) describe findings from Processing Instruction and C-BLI, (b) note why the theoretical underpinnings “leave something to be desired”, and (c) explain how usage-based models can fill the gap. With regard to (a), the discussion of actual findings is rather light—they are simply summarized in one sentence. With regards to (b) and (c), I also think there is some discussion missing. As my own work is in PI, not C-BLI, I’ll highlight what I see as the shortcomings of this discussion from that perspective, though I suspect someone with a background in C-BLI might have similar comments with reference to that paragraph: First, I’d note that in the discussion of PI, it is simply stated that the theories underpinning PI “lack explanatory power of underlying, biologically plausible mechanisms for the learning that occurs”. But here, there is nothing offered in the way of explaining why the accounts given in this work are insufficient (I might, for example, look at VanPatten & Rothman (2014) for a detailed account of the mechanism said to be at work). Second, the discussion glosses over the fact that work on PI has occasionally been discussed as theory-independent (despite VanPatten’s clear orientation towards UG frameworks), and some works have made explicit attempts to combine insights from the PI literature with usage-based psycholinguistic approaches (e.g., Henry, 2023). I’d highlight two quotes to illustrate the point: VanPatten (2015) says: “To be sure, I take a formal approach to the nature of the underlying system, but as I state in various publications, one need not take such a perspective to understand how PI might help language grow in the mind/brain (see, for example, VanPatten 2009, 2014).” With reference to VanPatten’s lexical processing principle, Dracos & Henry state: “Although the underlying theoretical assumptions are different, models of learned attention and cue blocking make similar predictions given that salient lexical and discourse cues typically overshadow verbal morphology in natural learning contexts (see, e.g., Ellis and Sagarra 2010a, 2010b)” (Dracos & Henry, 2021, p. 2). In general, I felt like this section didn’t “pay off” the way it could have. The ideas are here, and I do appreciate the explicit attempt to bridge the gaps between the PI / C-BLI worlds and functionalism / connectionism, but I think the conversation needs to be expanded in these ways in order to really make an impact.

Author Response

Added: citation for Henry, N. (2023). The additive use of prosody and morphosyntax in L2 German. Studies in Second Language Acquisition, 45(2), 348-369.

Added: According to VanPatten (2017), “PI is actually a type of focus on form or better yet, a pedagogical intervention. As such it is not a method with an underlying approach but instead an intervention that can be used by any communicative approach that seeks a supplemental or periodic focus on the formal features of language,” (p. 166). From this perspective, PI should be applicable under any number of language theories. However, in the same paper, VanPatten (2017) declares that “Under PI, acquisition consists of three necessary ingredients: input; Universal Grammar (UG) and internal mental architecture; and processing mechanisms that mediate between input and UG/internal architecture,” (p 167). Here, there is a sharp contrast to the theory-neutral stance of its applicability and a declaration of the need for UG to explain the learning mechanism. The major challenge here is that the definition of UG has changed significantly over the past decades (see Trettenbrein (2015) for an overview) , and our understanding of human cognition has developed to handle the difficulties that UG hoped to solve in more empirical ways.

Changed: Despite the UG foundations of VanPatten’s PI, it has been shown to be effective. Other researchers take up the more theory-neutral side of PI and have investigated it along lines that explore PIs connection to other theories of langauge, such as Henry (2025), who describes PI as having two “core tenets […]

Added: distances itself to some degree from leaves something to be desired regarding underly-ing, biologically plausible mechanisms for L2 learning, and does not necessarily disre-gard, but does not attend significantly enough to the properties we see in implicit and statistical learning that are still required for L2 development. The procedures involved in C-BLI are highly effective at helping students understand form-meaning connec-tions, notice forms, apply relevant higher-order thinking, and ultimately apply this in-formation, but it does not help us understand the route between this explicit infor-mation and automatization and/or proceduralization of this knowledge into implicit knowledge.

Reviewer Response

The additions here are helpful, though I do still think it would be relevant to at least footnote the 2015 quote from VanPatten where he states that “as I state in various publications, one need not take such a perspective to understand how PI might help language grow in the mind/brain”. The quote that you highlight might be a development in VanPatten’s thinking, but it seems to me to be more of an aberration. Either way, I think it’s relevant to note his earlier stance.

Author Response

The additions here are helpful, though I do still think it would be relevant to at least footnote the 2015 quote from VanPatten where he states that “as I state in various publications, one need not take such a perspective to understand how PI might help language grow in the mind/brain”. The quote that you highlight might be a development in VanPatten’s thinking, but it seems to me to be more of an aberration. Either way, I think it’s relevant to note his earlier stance.

Added footnote to include quote and citation.