Talk:Bootstrapping (linguistics)

I am going to bridge the gap between this page and the main Syntactic bootstrapping page; I will integrate the information leaving only a brief introduction on this page and moving all other content to the other page as Wiki does not want a lot of repeated content. I will report my updates here as I go. Olíe (talk) 01:17, 7 December 2017 (UTC)Olíe

Revisions

Following the suggestions and comments, we have expanded our information and have included more references as needed. We have also tried to change our wordings so that everything is more comprehensive to those who are not part of the Linguistic field of study. We have also changed some of the formant of the article, such as citations, examples and subheadings, so that hopefully it will be easier to navigate. Hopefully, the added information will continue to uphold the praise given with regards to factors such as clarity, neutrality, comprehensiveness.

Comments on content & structure

In addition to discussing the relation of bootstrapping to connectionism, also discuss it relation to the "innateness hypothesis". In the section on semantic bootstrapping, subdivide the discussion in terms of semantic classes: discuss how the state/event and the count/mass contrasts are acquired. For syntactic bootstrapping, subdivide the discussion int the acquisition of lexical versus functional categories. For the lexical categories, discuss how each of the four lexical categories are acquired, namely verbs, nouns, adjectives, and prepositions. For prosodic bootstrapping, subdivide the discussion in terms of the different "domains" that prosodic bootstrapping picks out: word-level, phrase-level, sentence-level, and utterance-level prosody. --RM Dechaine (talk) 17:14, 4 November 2014 (UTC)

Peer Review (A2)

Semantic Boostrapping

Hello Group A1! Nice start off and I hope you had fun in the first draft. The following is the suggestions and comments made based on the section of introduction and semantic bootstrapping:

(i) clarity When syntax is mentioned, it may be better to replace the word by grammar or sentence structure, so people who are not familiar with linguistics or syntax can still understand the relationship investigated. Assumption 3: Language addressed to a child must be accompanied by non-syntactic cues. (not sure what do you mean here; so a child needs non-syntactic cues to learn a language?) Assumption 4: Children must have some sort of innate knowledge of certain linguistic principles. For example, agents of transitive verbs (you may need to explain or link to another wiki page for readers, especially for those who are not familiar with linguistics; they may not know what is agent and/or transitive verbs) are the subject of the sentence (is there any case that the subject of a sentence is not agent of the transitive verb? If so, it may be better to say “most of the time” or “usually”, instead of “are” here)

(ii) verifiability; It's very nice to give examples to illustrate how the assumptions applied onto the real sentence and learning process. It would be more comprehensive if assumption 3 and 4 had been applied to example 1, so as to help readers to understand each of the assumptions being made by Pinker.

(iii) Neutrality; More different opinions can be suggested. Since in this section, it seems only Pinker, who first introduced semantic bootstrapping, is mentioned. The development after Pinker can be also suggested to strive for neutrality.

(iv) quality of research; Good job. The summary of the article is detailed and is easy to follow.

(v) comprehensiveness; In term of semantic, it might be better if negative example was provided too, for instance, how a speaker still get the meaning of the sentence when the sentence is grammatically ill-formed.

(vi) lay-out; Layout is fine here and the example being cited is placed in a correct box. However, should the sentence inside the box be italic?

(vii) citations; The Pinker part is fine but the part that introduced agent of transitive verb, object and noun should be cited.

(viii) Images; There is no image so no comment up to this point. But I don't think image here plays an important role, that is, image is not necessary here.

(ix) style; Style is fine and is close to typical Wiki style, but less usage of “that” might be easier for the readers to follow.

Waitingxu (talk)

Syntactic Bootstrapping

Hello A1, your article is coming together very nicely! I have looked at the section titled ‘Syntactic Bootstrapping’. This section is very clear and well supported. You have done an excellent job of staying neutral. It looks like you have done plenty of research in this area, and you have some really clear examples of studies done. It might be helpful to have a couple more examples. I see you have a great study on the acquisition of verbs, but can you find some information of the acquisition of other word classes? The lay-out of this section is straightforward, perhaps a few more subheadings as mentioned above would be helpful. The citations you have are good, I would suggest however, that you should add some more blue links into our article, in order to help those who are unfamiliar with linguistic terms. All in all, you’ve done a great job. Keep adding more information, studies, and examples. Perhaps adding some examples that are not english would be good too. Looking forward to seeing the final product! Halinapoho (talk) 23:28, 3 November 2014 (UTC)

Prosodic Bootstrapping

Hello! Good job on the page, I can tell you've put a lot of research and effort into it. I think the prosody section is a pretty clear one, which links to other wikis for terms that might be confusing. However, there needs to be more of these, as there are a lot of Linguistic terms that someone that knows little of Linguistics might not know. For example, you should link back to terms such as stress/rhythm, since it is not explicitly stated what these are. Also with intonation/pitch/accent, since these could be ambiguous terms. Other sections that I would recommend to clear up/link to another wiki are CDS/IDS, paralinguistic, unbound morphemes, and auxiliary verbs. I think you've done a good job with making it sound neutral and informative, and the diction is very comprehensive. The layout is also very neat and clear. However, there are some sections that would be more credible with the help of a citation. For example, the part about prosodic envelope, the end of CDS section, and the criticism section. Dianamen (talk) 06:14, 4 November 2014 (UTC)

Pragmatic Bootstrapping

Hello A2! I enjoyed reading through your article, and found the information on "gaze following" interesting and informative. Keep up the good work on expanding your content! The following are just a few suggestions and evaluations of the Pragmatic Bootstrapping section:

(i) clarity; The text is clear and nowhere is anything which is communicated opaque or hard to discern.

(ii) verifiability; There is no original research included, and everything which is discussed is cited, or at least refers to its source content so that its authenticity can be verified.

(iii) neutrality; The tone of this article is neutral.

(iv) quality of research; The research is well done, but more content could be included to give the reader a greater and broader introduction to the topic.

(v) comprehensiveness; The article is comprehensive and easy to understand.

(vi) lay-out; The layout is overall quite good. However, the text in the two boxes labled “Action” and “Object Highlighted Condition,” respectively, are formatted such that they run along in a single line for the entirety of their content. It would be best to re-format these so that one does not have to scroll so far to the right to read it.

(vii) citations; Good citations are included, however, the text after “experiment by Michael Tomasello and colleagues” requires a citation.

(viii) images; An image of parents speaking to children, or something illustrating “gaze following”, may be good to add.

(ix) style The style is well done and retains an air of neutrality and generality suitable for an encyclopedia article.

Dbell92 (talk) 04:54, 4 November 2014 (UTC)

Annotated Biblography

Siklossy, L. (1976). Problem-solving approach to first language acquisition. Annals of the New York Academy of Science, 280, 257-261. The article outlines a computational theory regarding language acquisition. Siklossy hypothesises that children use various problem-solving techniques, similar to that of computer programs, in order to sophisticatedly learn a language. Great stress is placed upon the ability of a child to pull relevant pieces of information from a string of speech, and then construct mental representations of the concept. Using these early representations, the child is then able to use this information in order to decipher other unknown elements of speech. However, whereas computers are able to perform these basic functions, they lack the sophistication of children’s cognition, which allows the encoding of detail and other complex grammatical functions. Considering that bootstrapping experiments did not emerge until the 1990’s, this article provides information with regard to the origins of the theory. The author is positing a connectionist perspective, where the computer-like functions of the brain are stressed, allowing one to see how the theory of bootstrapping emerged and was subsequently shaped by various different viewpoints. The article does however lack in that it is purely a theory, and does not provide any subsequent experimental evidence, aside from comparisons to computer programs. This, however, has to do with the lack of information regarding the topic at the time, and, as mentioned, demonstrates an early perspective. The article provides interesting background with regard to the theory, and opens up the viewpoint of language acquisition as being a function of the brain’s statistical inferences and abilities. Despite its lack of experimental evidence, the viewpoint of this article should still be considered when discussing the topic of bootstrapping.

Kiss, G.R. (1973). Grammatical word classes: a learning process and its simulation. Psychology of Learning and Motivation, 7, 1-39. This article also approaches language acquisition and processing from a connectionist perspective. Kiss describes word classes in terms of linking theory, wherein each word has a separate internal representation and possesses various transitional links between it and other concepts. Neighbouring words thus provide the concepts and links for children to engage in bootstrapping by using their previous knowledge to create new representations. The article heavily stresses the computational aspect of language acquisition, using computer programs as demonstrations of linguistic capability. However, despite the stress on connectionism, the article utilises these simulations in a beneficial way, in that it aims to demonstrate the logical and associative aspects of cognition involved in language acquisition. One advantage of this article is that it utilises empirical evidence by assessing the lexicons of one and three years, comparing similarities and differences across all. One issue with the article however, is it places large emphasis on the human brain as being able to be broken down into algorithms and formulas. The bias here creates an issue in that it does not explain the complex elements of speech that are easily recognized by humans, but not by computer algorithms, such as prosody or intonation. The value of this article ultimately is that it provides the connectionist background behind the theory of bootstrapping. Despite its high dependency on computational simulation, the article’s value is in the fact that it demonstrates the early theories of statistical learning and bootstrapping.

Hoff-Ginsberg, Erica & Shatz, Marilyn. (1982). Linguistic Input and the Child’s Acquisition of Language. Psychological Bulletin, 92(1), 3-26. Hoff-Ginsberg and Shatz present a clear and broad academic literature review on language acquisition. They focus on semantic and syntactic development, models of acquisition, and briefly summarize previous research findings. As the article has good breadth, it only briefly touches on the concept of bootstrapping. Because bootstrapping was only formally introduced by Pinker in 1981, it is acceptable that it lacks a depth of discussion. This article provides good insight into the history of language acquisition and early research. It introduces bootstrapping in the context of current (1980) research and findings, which may be of interest if our Wikipedia group chooses to discuss the history of bootstrapping. Hoff-Ginsberg and Shatz do not formally side with a particular acquisition theory but rather present a general overview of information. They examine the debate of language being innate, universal grammar, and the interaction between both internal and external factors contributing to language acquisition. Additionally, they briefly discuss language modelling and language acquisition for deaf children. Hoff-Ginsberg and Shatz do a good job of presenting many aspects and theories on language acquisition and do not strongly side with any of the perspectives. They do, however, suggest that there may be an interaction between nature and nurture that contributes to the acquisition of language. I do believe that this article may be of value to our Wikipedia group as it shows the development of language acquisition research over the years and is a good contrast to more current research. It provides a good examination of multiple aspects and theories, but overall does not provide a great deal of information on bootstrapping. As a result, this article may be good for composing a history of language acquisition section but not as a main source of information.

Bohannon, J. N., & Leubecker, A. W. (1988). Recent developments in speech to children: We've come a long way, baby-talk. Language Sciences, 10(1), 89-110. doi:10.1016/0388-0001(88)90007-1

In this article Bohannon and Leubecker examine how children intake language information. They propose that children acquire language by listening to every-day conversations, and attempt to find out just how much and how rich the information they acquire is. This is relative to Bootstrapping as it discusses how children acquire language by baby talk and how parents speak to the child that adds to its learning environment. They not only took into account the contradictions, but also of information that they were not able to acquire, such as “what mothers are not saying” (100). It was interesting that the research not only account for children, but also for the parents and their learning environments.

Leititia R. Naigles and Erika Hoff-Ginsberg. (1998). Why are some verbs learned before other verbs? Effects of input frequency and structure of children’s early verb use. Journal of Child Language, 25, pp 95-120 This paper explores the topic of verb acquisition in children. The researches Naigles and Hoff-Ginsberg conducted a study regarding the correlation between the frequency children heard certain verbs, and how quickly they learned them. Naigles and Hoff-Ginsberg predicted that the more frequently a verb was heard, the more quickly the child would learn it. Although this study is the first on this particular topic, other similar studies regarding language acquisition in children have been done previously. Brown (1958) conducted research regarding the acquisition of nouns, and Gentner (1982) studied whether nouns were learnt before verbs. This study was conducted by recording a child (who was just beginning language acquisition) speaking with their mother in various settings and counting the number of times and in which contexts certain verbs were said (by the mother). The context, or “frame” of the verb was an important aspect to test for bootstrapping. Then, 10 weeks later, the child was brought back to assess whether the verbs spoken most by the mother were indeed the verbs the child learnt first. It was found that this correlation between frequency of input and frequency of use 10 weeks later was strong. From this data, it is clear that children pay attention to the context in which verbs are said in order to internalize meaning, and this is demonstrated well in the research through the use of many subjects to get as much data as possible, and the thorough way in which the data was analyzed. In conclusion, it is a thorough study that gives some valuable information on child language acquisition. Janna B. Oetting. (1999). Children with SLI use argument structure cues to learn verbs. Journal of Speech, Language and Hearing Research, 42, pp. 1261-1274 In this article, Oetting examines whether children diagnosed as specifically language impaired (SLI) use bootstrapping to the same extent as other children when learning verbs. Previous studies (Lely 1994, O’Hara and Johnston 1997) have found that children with SLI have difficulty processing the structural aspects of language. However, other studies, such as the one done by Hoff, Ginsberg, Kelly and Buhr in 1996 seem to suggest that children with SLI did in fact understand the grammatical features of sentences. The study was conducted by engaging the children in two tasks. One was giving them a nonce verb in a sentence, and then having them watch a video clip of two characters completing an action, to see if they would make the correlation between the two. This was to examine bootstrapping in the children. In the second task, children watched a videotaped story containing a sentence with a nonce verb in a transitive or intransitive sentence, correlated with an action scene. After three viewings of the tape, he children demonstrated their understanding of the correlation by pointing at pictures portraying the action when the word was heard. This task was to examine verb retention in the children. Oetting found that children with SLI were able to gather the meaning of verbs from cues. It was also found that verb retention was lower in the SLI children than the controls. However, the SLI children did show an understanding that the made up words were verbs, as they did not point at the pictures of nouns. This study demonstrates the difficulties that SLI children have with acquiring verbs. However, it was found that this difficulty comes from retention and not a lack of understanding of the grammatical structures. This is a valuable insight when teaching children with disordered speech.

Lee, Joanne N & Naigles, Letitia R. (2008). Mandarin Learners Use of Syntactic Bootstrapping in Verb Acquisition. Cognition, 106. This article examines whether Mandarin learners extend verb meanings based on the number of noun phrases in a sentence. In studying two and three year old Mandarin learners, researchers concluded that the number of noun phrases in a sentence appeared to play a role in Mandarin language learning. This article is significant to our Wikipedia project as it provides a brief overview of the syntactic bootstrapping theory and provides in depth cross-linguistic findings about syntactic bootstrapping. Similarly, the study discusses the finding’s implications in comparison to English language acquisition and briefly considers the innateness of using argument number to infer sentence meaning. This information is essential to our project when discussing bootstrapping cross-linguistically and relating it back to language acquisition as a whole. This is an academic article, written from an academic and research perspective. It clearly concludes study finding, discusses implications, and considers shortfalls and third variables. While this paper is not biased, the sample of participants was very small and was not entirely representative of all Mandarin speakers. As a result, the findings of the study cannot be generalized to the entire Mandarin speaking population. Overall, I believe this article is a good starting point for our Wikipedia project, provides good cross-linguistic findings, but does not offer enough sustenance on syntactic bootstrapping as a whole.

Gray, Wayne & Schunn, Christian. (2002). Bootstrapping in Miniature Language Acquisition. Cognitive Systems Research, 3. This article explores the relationship between semantic and syntactic bootstrapping and how the two enable language learners to understand the meaning of words and phrases. It questions how children learn the phenomenon of bootstrapping and the role of a connectionist network in language acquisition. I believe this article has potential to be of value to our Wikipedia project because it provides an excellent overview of syntactic and semantic bootstrapping. It discusses with good depth how bootstrapping works and how it acts as a learning tool. This article, however, has some limitations. While it considers multiple perspectives, the article heavily focusses on the role of the connectionist network in language acquisition and is not a research paper about a study, a literature review of many studies, or a proposal for a study. As a result, I think this is a good summary of key facts about bootstrapping. It includes many citations of credible papers, researchers, and sources that our group can consult for further reading and information.

Arias-Trejo, N., & Alva, E. A. (2013). Early spanish grammatical gender bootstrapping: Learning nouns through adjectives. Developmental Psychology, 49(7), 1308-1314. doi:10.1037/a0029621

Arias-Trejo & Alva examines how grammatical gender and numbers may affect Spanish-learning toddlers with regards to learning nouns by knowing/understanding the masculine/feminine adjectives. They argue that these toddlers benefit from learning grammatical gender and numbers to understand the relationship between words and what they refer to. My Wikipedia project topic is bootstrapping, where we are interested in how children acquire language. This article is significant because it helps us to understand how toddlers are able to understand the relationship between a word and the object it refers to, and that such learning is “flexible” (1312). It also provides evidence of how in Spanish despite how complex masculine and feminine adjectives are, they instead benefit toddlers in acquiring this language relationship. They mention that this study further proves the current research that is also studying language acquisition through different grammar styles, and so perhaps this article was written to simply further prove instead of possibly looking further for points that may disprove this research. I found this research to be quite interesting, because I have had the assumption that masculinity and femininity in languages would make language acquisition much more difficult (as I experienced this when learning French).

Gentner, D., & Christie, S. (2010). Mutual bootstrapping between language and analogical processing. Language and Cognition, 2(2), 261-283. doi:10.1515/langcog.2010.011 In this article, Gentner examines our relational cognitive capacity for language acquisition, as well as how we are able to apply this cognitive knowledge and use it in everyday language. They also argue that the method of acquiring language would be where humans learn through experience, rather than innately, and this experience is used to compare with other experiences. This is relative to Bootstrapping because it explains and provides evidence for different methods of language acquisition. This article provides a comparison between itself and other theories/articles, and so I don’t think there is much, if any, obvious notable points of bias. Although I personally prefer the Innateness theory, it was interesting to read about how language acquisition and experienced worked together. — Preceding unsigned comment added by Jcjjfu52 (talk • contribs) 06:26, 3 October 2014 (UTC)

Comments

This needs a lot of work. 128.158.145.51 14:42, 18 July 2007 (UTC) RM Dechaine (talk) 05:23, 1 September 2014 (UTC)

Suggestions for final round of revisions

In its current form, this is a "text-heavy" entry. Set off the examples from the main text, and introduce more visual material. Also note that some references are missing. The "finger pointing" section has no content, so should perhaps be deleted. The content of the "semantic bootstrapping section" needs to be extensively re-written, and more clearly connected to the main point of the article.--RM Dechaine (talk) 17:48, 7 December 2014 (UTC)