Talk:Negative evidence in language acquisition

Wiki Education Foundation-supported course assignment

This article was the subject of a Wiki Education Foundation-supported course assignment, between 27 August 2019 and 11 December 2019. Further details are available on the course page. Student editor(s): Coffeekoala. Peer reviewers: Zhujialei, Hyeonah, RemoLing.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 01:30, 18 January 2022 (UTC)

Indirect Negative Evidence

These sections of the article are somewhat confused. I agree with Lord2019 that constructivism and nativism subsections could be useful. I don't currently have enough information to flesh out these subsections, so for now, I'll reorganize the Indirect Negative Evidence section to make more sense. Proposed order:

    Indirect Negative Evidence in Language Acquisition
        (give lang acquisition as well as general definition here)
        Utility of Indirect Negative Evidence
            Word Learning
            Syntax
                (paragraph about verbs should actually be here. It's about learning what syntactic frames verbs fit into rather than the actual meaning of disappear/vanish)

Coffeekoala (talk) 18:22, 15 October 2019 (UTC)

Notes for Future Drafts

Distinguish sections as “Indirect Negative Evidence in Constructivism” and “Indirect Negative Evidence in Nativism” as opposed to “early/recent research” Lord2019 (talk) 17:34, 20 October 2017 (UTC)

Peer Review

In Indirect Negative Evidence in Language Acquisition, you need to fix the citation from Lust's textbook (make it a citation that has a reference on the Wikipedia page). Including actual examples of indirect negative evidence could also be helpful in explaining what it is. I think the definition you give is a little unclear (what exactly is not heard in the language input?). I also thought that the evidence that you are citing in the early research section is a type of implicit (not indirect) negative evidence since it is present in the speech input (but I might be totally wrong on that--it just might be worth looking into to make sure. I think Lust talks a little bit about the two different types of evidence at the beginning of ch 3, so it might be helpful to look there).Rcsender (talk) 19:52, 31 October 2017 (UTC)

The structure so far looks good. This is still a rough draft, of course, but hopefully you will go into further detail here: "Much earlier research supports a nativist view (most prominently, Marcus 1993[2]). More recent research adds further support to this view." Cgilchri (talk) 03:41, 7 November 2017 (UTC)

This is a pretty good start, but you still have a ways to go. Here are some comments that I hope are helpful. 1) Please make sure to distinguish implicit negative evidence (which is discussed by another group) and indirect negative evidence (which is what you guys are supposed to talk about). Implicit negative evidence is a kind of parental correction. Indirect negative evidence is the idea that a child could use the absence of a given property as evidence that that property is absent. This kind of inference is not just about language, but about any kind of inference from evidence. For example, can I conclude that the sun will rise tomorrow, given that it has risen every other day? Technically, this inference is not warranted just from the previous data, since maybe the sun skips a day every 200 years. But, given that I have not seen (or heard of) a day when the sun didn't rise, I might be able to be more confident that it will rise than that it won't. This kind of inference to the best explanation is a kind of reasoning by indirect negative evidence. See the discussion in Regier and Gahl 2003, which I think is helpful. I thnk the discussion in Lasnik & Lidz (that we read for class) might also help clarify this. (2) One place where indirect negative evidence has been argued to be useful is in word learning (see Xu & Tenenbaum 2007 for a good example). Also Regier and Gahl (2003) argue for indirect negative evidence in learning one-anaphora (but see Pearl & Lidz 2009 for a rebuttal). You can also find some relevant stuff in work by Anne Hsu and by Ben Ambridge (he probably calls it something like entrenchment). (3) I would try to take out the nativist/constructivist discussion. I don't think that distinction is entirely relevant here (though many of the researchers involved probably frame things that way). If inference from indirect negative evidence is a kind of reasoning that humans can do, then the nativist's job is to explain where the relevant hypotheses come from and the constructivist's job is to explain how these hypotheses can be constructed. So, just stick to the question of (a) what it is (b) whether there are phenomena for which it would be useful, and (c) whether child language learners have been shown to use it. I hope this helps. JeffLidz (talk) 16:57, 8 November 2017 (UTC)

For the most part your information seems good. However, the style could use a little help. Some sentences could be restructured to read more fluidly. For example: "The theory in regards to language acquisition is..." re word to "In regards to language acquisition, the theory (what theory..?) states that..." I would also go through and check for parallel structure, agreement, and punctuation. Maggicurrier (talk) 04:07, 17 November 2017 (UTC)

I like how citations and other Wikipedia pages are linked. I have not seen Wikipedia articles divide up section headings for early/late research, so maybe something like "history" would be more apt. There is a lot of discussion about nativism intertwined with other ideas, so maybe it would be better to make the separation of concepts clearer. Jchung10 (talk) 21:41, 17 November 2017 (UTC)

Thanks for your feedback! We did a lot of restructuring regarding the layout and I think it makes more sense. Take a look and let me know if you agree, and if it is understandable. 108.51.207.41 (talk) 16:31, 19 December 2017 (UTC)

Updates

Just edited the page and am redoing the sections. I deleted the section on Early Research and support for nativism/constructivism. I know the article looks insanely short at the moment but I will be adding more soon. There is much more to be added so stay posted Lord2019 (talk) 14:55, 28 November 2017 (UTC)

Hey! Just added the direct negative evidence material that my group has put together... Still need to add citations, but I will be working on that soon! Rcsender (talk) 18:31, 8 December 2017 (UTC)

Comments on Things to Clarify Before Final Article

Hey all! I’ve gone through the whole article for general grammar/flow things, and fixed all the mistakes I spotted. I think during our meeting, we can discuss how/if we want to reformat anything so that the page seems more coherent and cohesive after all. I haven’t changed any informational things, but there were some things I think we might want to clarify:

• Overall, we should retitle the page “Negative evidence in language acquisition,” and then develop a general intro paragraph (it can be pretty brief) that states that there are two main types, Direct and Indirect. Then we can launch into the rest of the article.
• Added a link to the “grammaticality” Wiki page on first time we mention grammaticality; do we think this is necessary/helpful?
• In “Do children use negative evidence”
  • There’s a sentence that says “experiments show that children produce more grammatical forms…” “More” is ambiguous here. Are kids producing a larger number/proportion of grammatical sentences, or sentences which have a higher degree of grammaticality (on average, or than previously).
  • “Children tend to directly respond to these reformulations…” What’s the referent of “these?” Parental reformulations?
  • On the study of the mute child: How did he produce utterances (which are spoken) for the study if he was mute? Did he become un-mute, or were they having him write sentences down?
• In “Direct negative evidence in language learning”:
  • The first sentence is quite long, and thus hard to follow. It should probably be revised for clarity.
  • On the word “Nativism,” added link to the “Language” section of the “Psychological Nativism” article. Do we want to link to the UG article instead, or no?
  • Should we move the “Examples in Language Acquisition” section for INE into the Syntax section of “Utility of Indirect Negative Evidence”? We should also very briefly describe Principle C and anaphoric one, as a naive audience (e.g. students in an intro-level ling class) might not know what those things are
• In “Indirect Negative Evidence and Word Learning”:
  • What ages were the participants in Xu and Tenenbaum’s study? Feels like it would be useful to say something like “X-month-old participants,” to hammer home the point that it’s young learners acquiring language
  • We need a citation(s) for the disappear/vanish section; it could also be edited for clarity (especially the last sentence with the “that that” construction)
• In “opposing viewpoints”
  • We still need to insert a citation where it says “insert a specific citation”
  • Should we clarify what the “U-shaped curve” is? Again, I feel like naive audiences wouldn’t be familiar with this.
• It also feels like we could have more examples/discussion of INE in general
• Some of the formatting on the citations is wonky, but I’m not sure how to fix the citation formatting. We can look into it

Shkaplow (talk) 06:17, 15 December 2017 (UTC)

Miscellaneous from the article

Layout Proposal:

• Negative Evidence (Linguistics)
  • Direct
    • ...
  • Indirect
    • Background
      • Frequency in input
      • Children response to INE
      • INE in L2
      • Learning algorithms, INE role in arguments against nativism
    • Implications for LA
      • Poverty argument
    • Applications/Examples

Opposing viewpoints

Noisy feedback is a type of input children receive from adults that includes ungrammatical utterances, speech errors and fragments, providing conflicting evidence for children during acquisition. Some have argued that indirect negative evidence guides children to the correct grammar due to the lack of ungrammatical utterances in their input, but noisy feedback would interfere with correct, grammatical feedback (insert a specific citation). Marcus, however, argues that noisy feedback is ignored by children during language acquisition.

Another argument against the usefulness of indirect negative evidence is found in Marcus et al’s ‘Overregularization in Acquisition,’ in which children show an inexplicable U-shaped curve when learning English irregular past tense forms. Young children are shown to very rarely overgeneralize in the English past tense (e.g. say ‘eated’ instead of ‘ate’), yet slightly older children do overregularize, and even older children eventually develop adult grammars without these overregularizations. Throughout these stages, though, children receive the same input from parents and adult caregivers, without a change in indirect negative evidence.

Possible Sources

• Lust, Barbara. Child Language: Acquisition and Growth, Cambridge University Press, 2006. [Ch. 3]
• Gathercole, Virginia C. Mueller, and Erika Hoff. “Input and the Acquisition of Language: Three Questions.” Blackwell Handbook of Language Development, 2007, pp. 107–127., doi:10.1002/9780470757833.ch6.
• Lidz, Jeffrey, et al. “What Infants Know about Syntax but Couldn't Have Learned: Experimental Evidence for Syntactic Structure at 18 Months.” Cognition, vol. 89, no. 3, 2003, pp. 295–303., doi:10.1016/s0010-0277(03)00116-1.
• Regier, Terry, and Susanne Gahl. “Learning the Unlearnable: the Role of Missing Evidence.” Cognition, vol. 93, no. 2, 2004, pp. 147–155., doi:10.1016/j.cognition.2003.12.003.
• Marcus, Gary F. “Negative Evidence in Language Acquisition.” Cognition, vol. 46, no. 1, 1993, pp. 53–85., doi:10.1016/0010-0277(93)90022-n.
• Pinker, S. (1989). Learnability and Cognition: The Acquisition of Argument Structure . Cambridge, MA: MIT Press. [Ch. 1]
• Bohannon, J. N., & Stanowicz, L. B. (1988). The issue of negative evidence: Adult responses to children's language errors. Developmental Psychology, 24(5), 684-689. -- Applications/Frequency
• Farrar, M. J. (1992). Negative evidence and grammatical morpheme acquisition. Developmental Psychology, 28(1), 90-98.
• Clark, Alexander, and Shalom Lappin. “Another Look at Indirect Negative Evidence.” Proceedings of the EACL 2009 Workshop on Cognitive Aspects of Computational Language Acquisition - CACLA '09, 2009, doi:10.3115/1572461.1572467.
• White, Lydia. “Adverb Placement in Second Language Acquisition: Some Effects of Positive and Negative Evidence in the Classroom.” Interlanguage Studies Bulletin (Utrecht), vol. 7, no. 2, 1991, pp. 133–161., doi:10.1177/026765839100700205. -- Applications/Frequency
• Izumi, Shinichi, and Usha Lakshmanan. “Learnability, Negative Evidence and the L2 Acquisition of the English Passive.” Second Language Research, vol. 14, no. 1, 1998, pp. 62–101. JSTOR, JSTOR, www.jstor.org/stable/43104559. -- L2
• Brooks, Patricia J., and Michael Tomasello. “How Children Constrain Their Argument Structure Constructions.” Language, vol. 75, no. 4, 1999, pp. 720–738. JSTOR, JSTOR, www.jstor.org/stable/417731. -- Opposing argument
• Xu, F., & Tenenbaum, J. B. (2007). Word learning as Bayesian inference. Psychological Review, 114(2), 245-272.

http://dx.doi.org/10.1037/0033-295X.114.2.245

• Lasnik, Howard. (1989). On Certain Substitutes for Negative Data. 20. . 10.1007/978-94-009-0955-7_5.

Allexan (talk) 18:06, 5 October 2017 (UTC)

Sources before 2000:

Rohde and Plaut 1999. “Language Acquisition in the Absence of Explicit Negative Evidence: How Important is Starting Small?”: https://pdfs.semanticscholar.org/e712/e8f8df025c8d2df45c4b08b5fbee117e15ae.pdf

Marcus 1993: “Negative evidence in language acquisition”: http://www.atsweb.neu.edu/hlittlefield/CourseDocs/Acq/Marcus-1993-pt1.pdf

Bohannon & Stanowicz 1988. “The issue of negative evidence: Adult responses to children's language errors.”: http://web.a.ebscohost.com/ehost/pdfviewer/pdfviewer?vid=1&sid=bf78c670-5f8e-41f1-b9ea-5eb8cce24bff%40sessionmgr4009 You may have to be on campus or sign in with UID to access this one.

Seidenberg 1997. “Language Acquisition and Use: Learning and Applying Probabilistic Constraints.” http://science.sciencemag.org/content/275/5306/1599.full

Bowerman 1988. “The ‘No-Negative Evidence’ Problem: How Do Children Avoid Constructing an Overly General Grammar?” http://pubman.mpdl.mpg.de/pubman/item/escidoc:468143/component/escidoc:532427/bowerman_1988_The-No.pdf

Sources after 2000:

Some rando’s summary, which is actually pretty useful as a jumping-off point for sources: http://www.academia.edu/30999674/The_role_of_negative_evidence_in_language_acquisition_current_state_of_research

Hsu & Griffiths 2016. “Sampling Assumptions Affect Use of Indirect Negative Evidence in Language Learning”: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0156597

MacWhinney 2004. “A Multiple Process Solution to the Logical Problem of Language Acquisition.”: https://www.cambridge.org/core/services/aop-cambridge-core/content/view/79F2699EA5184EB06CB19F1820680D67/S0305000904006336a.pdf/multiple_process_solution_to_the_logical_problem_of_language_acquisition.pdf

Ramscarr & Yarlett 2007. “Linguistic Self-Correction in the Absence of Feedback: A New Approach to the Logical Problem of Language Acquisition”: http://onlinelibrary.wiley.com/doi/10.1080/03640210701703576/full

Reali and Christiansen 2005. “Uncovering the Richness of the Stimulus: Structure Dependence and Indirect Statistical Evidence.”: http://onlinelibrary.wiley.com/doi/10.1207/s15516709cog0000_28/full

Clark & Lappin 2009. “Another look at indirect negative evidence.” Page 26-33 of the Proceedings of the EACL 2009 Workshop on Cognitive Aspects of Computational Language Acquisition. Very comp-sci heavy, so may not be useful: https://www.aclweb.org/anthology/W/W09/W09-09.pdf#page=38 Lord2019 (talk) 17:32, 20 October 2017 (UTC) 15:12, 15 December 2017 (UTC) — Preceding unsigned comment added by 129.2.181.227 (talk)

Comments by Jeff: 1) I don't think you can say in your introduction that Negative Evidence helps language learners acquire their language. That is a very contentious hypothesis. What you can say is that negative evidence is evidence about what sentences are not possible in the language. You can then say that it is controversial whether (a) negative evidence is available to children and (b) if it is, whether children use it to acquire their language. JeffLidz (talk) 15:51, 16 December 2017 (UTC)

2) Maybe add a link to Gary Marcus's wikipedia page? JeffLidz (talk) 15:54, 16 December 2017 (UTC)

3) Maybe add a sentence around citation 13 where you say that that study found that implicit negative evidence was a negative predictor of the rate at which children eliminated ungrammatical forms from their speech. JeffLidz (talk) 15:57, 16 December 2017 (UTC)

4) Please add more detail in the examples of indirect negative evidence section. I don't think someone unfamiliar with the topic could understand this paragraph. JeffLidz (talk) 16:00, 16 December 2017 (UTC)

5) I can see what you're getting at with the disappear/vanish example but please try to make it a little more explicit. You should say that because "disappear" is more frequent than "vanish", then the fact that it doesn't occur in transitive clauses (She disappeared her head) is more obvious than the fact that "vanish" doesn't. Learners can use the frequency of a verb to determine expectations that it might occur in other frames. If a verb does not occur in a particular frame, then the more frequent the verb is, the more suspicious that absence becomes, allowing learners to reach the conclusion that the verb cannot occur in that frame. JeffLidz (talk) 16:04, 16 December 2017 (UTC)

6) Try to give more detail about the nature of the argument that Regier and Gahl made. What is the indirect negative evidence and what conclusion do learners draw from that evidence? JeffLidz (talk) 16:06, 16 December 2017 (UTC)