From: Kathleen Currie Hall 1. Two basic models of phonological development: developmental/biological and cognitive/creative. In this paper, Vihman assumes a modified combination of the two -- biological foundations, with individual biological variation and limits on variation, influenced by the ambient language and shaped by active cognitive integration. 2. The focus of the paper is variation across phonological acquisition: there is evidence of some universal development; some language-particular development; and some idiosyncratic, child-specific development. Child-specific variation seems to come mostly from internal sources (anatomy, physiology, attention, memory, etc.). 3. The input that a child receives, the attention that he pays to that input, and the ability of the child to match up input with particular babbling that he has made himself (the "articulatory filter") are crucial factors in shaping phonological development. Questions: 1. Am I just being hypercritical? It seems like a lot of the conclusions / main points / etc. are based on very limited data and are not really backed up with solid evidence or statistics. (e.g. standard deviations in Tables I and II do not always "drop" as drastically as the text makes it seem; Vihman says says that internal sources of variation must dominate over external ones because the evidence "seems sufficient" -- based on what? etc.) Perhaps I am just not well-enough versed in this literature to understand what she was talking about, but I often got confused while reading this paper. 2. Can we look at more about how kids "develop a network of related forms" -- alluded to briefly on p. 78? From: Junko Davis What constitutes the "ambient language"? -- The article mentioned some recent studies that have reported language specific patterns in the pre-linguistic period (p.62) which indicate the significant influence of the ambient language. However, if the primary caretaker is not a major source of the individual differences in the children's phonetic production, what happens when the primary caretaker(s or entire family) speak a different language from the one that surrounds them? For instance, the language spoken in a family is Japanese, but the family lives in the United States. They tend to leave the TV on all the time, and the TV programs are all in English. Can TV be a source of input for the child in this case? What is the ambient language of this child? Is there any touchstone to determine if a developmental pattern can be "abnormal" rather than idiosyncratic variation? From: ASIMINA SYRIKA and LAURA SLOCUM 1. What exactly does Vihman mean by 'controlled expansion' and 'crystallization' (page 68)? She implies that these are similar to a child's learning style, but what influences whether a child is no 1 or no 2? And does no 1 imply typical development and the other one atypical? 2. Why does Vihman include 'name shape' under internal and not external factors in her discussion of sources of variation? How does she define 'child internal sources'? 3. Vihman quotes Lindblom: "child's speech production appears at first to be largely, if not entirely, system-oriented." (page 73) What does he mean by 'at first'? We recognize that children at young ages are motivated to have their needs understood by others and question the timing of Lindblom's observation. From: Kirk Baker Vihman (1993). This paper focuses on variability in consonant production for children learning to talk. vihman suggests that variability in production is related to a flexible process of phonological construction. My reading was either too critical or too cavalier because I didn't come away with a strong feeling of demonstration of how certain types of variability actually lead to particular strategies of phonological acquisition. My overall impression is that there is variability in both phonetic production and phonological organization, but not real clear on how one predicts the other. Questions. p.64 "when the adult frequency is only moderate". I wasn't sure what corpus this is being measured over, or what phonological unit is being measured. p.67 "early vocables or protowords...which may lack an adult model". I was surprised that even these early utterances lack an adult model, and wondered if transition frequencies, intonation pattern, etc. really show no correlation with the adult language? p.68 "initial consonant of adult word...most salient aspect of adult speech stream". This surprised me because based on talking to Anton, I had the impression that stressed syllables were the most salient thing in the speech stream for kids to latch on to...maybe she's talking about the same thing depending on the characteristics of adult words she's looking at. p.72 Wasn't really sure what to make of the discussion of Laurent organizing his phonology around /l/ without knowing the other kids' names and how they organized their phonologies. p.73 "will readily allow a misperceived word to lead a conversation in an otherwise unmotivated direction". This seems really interesting to me, I hadn't even considered that kids' early learning would be so strongly bottom- up, wonder if it implies statistical learning all the way up to pragmatics? p.78 Conclusion. I'm just curious as to the range of variation found in some presumably innnate behavior like the transition from crawling to running for kids, or learning how to hunt in cats. How much would variability surprise us there, would the same issues arise when trying to frame it in terms of "flexible construction" of a system? From: ejkong@ling.ohio-state.edu - Infants takes within-/across differences(variable paths) while they approach the phonological organization. - Which path for infants to follow is in part influenced by internal sources such as the psysiological maturation differences and the ability of integrating linguistic stimuli. - 'Articulatory filter' makes infants relate phonetic gestures with acoustic patterns, which eventually helps them with the word production by establishing links between adults' acoustic patterns (that matchs infants' motor schemes) and the situations (where those singals are used). - Then, in this model, between-infant-variability (i.e different paths) are caused by differnt 'perceptual salience', while within-child-variability reflects the learning process of word production using vocal motor schemes and filtering. question: - Can we get detailed descriptions to illustrate/contrast 'controlled expansion' and 'crystalization' in p68? From: Helena Riha These are some thoughts I had in reading Vihman (1993). p. 62 "high individual variation across children early in the period of transition to speech can be accommodated iwthin a biological model": The possibility of individual variation in language acquisition would allow for more flexibility in responding to different caretaking styles and language input across cultures (although Vihman later notes that mothers are not a major source of individual differences in children's phonetic productions [p. 69]). p. 63 "Within this framework of universal constraints and language-particular tendencies ... we see considerable variation across individual children": Individual variability in early language acquisition may benefit humans in developing language because children need to come into the world ready to acquire any language. The fact that language-particular tendencies can be quite variable across languages means that human children need to be flexible when initially acquiring language. Their task contrasts with that of other species (other primates, for example) which acquire a closed communication system with a limited set of calls. p. 73 "the child's speech production appears at first to be largely ... system-oriented": This may reflect the child's evolving awareness of the self and others. As children matures cognitively, they also gain an increased awareness of and empathy toward others. p. 73 "the child must develop an understanding of ... routine situations": This may contribute to understanding why linguistic routines (greetings, etc.) are important in communication. p. 77 "It is the integration of the 'notion of [vocal] communication ... with the 'notions' of language-like sound shapes ...which appears to be critical for the emergence of language-like word use.": According to Papousek & Papousek, children integrate the notion of "vocal communication" with their mother's intonation early on. They understand the meaning of a mother's intonation contours even if they do not yet understand her words. This should be taken into account in understanding how children come to link the "notion of vocal communication" with "language-like sounds." From: Fangfang Li This paper investigates individual variations in early stages of phonological acquisition. The kind of "uniformity" at early stage of phonological acquisition as claimed by Lock (1983) still gets recognized, and is interpreted as the result of "development", which "reduces individual variability". Firstly, those pre-linguistic individual differences in vocal production are attributed largely to "biological" differences within languages, and an effect from ambient languages is attested cross-linguistically. Also, with a certain language, those individual differences are more from the internal differences than from the external ones like the input from the primary care givers. It points out the these type of variation has significant implication for child language acquisition since it helps children to map the articulatory gestures with their perceptions through exploring the vocal space. Questions: 1. The paper mentions that the articulatory filter is "unique to each child'. Then, will children speaking the same language will share more commonalities in their aritulatory filter than them with children speaking a different language? In other words, if this articulatory filter is unique to each individual, then how the uniformity or universality can be explained in this term? 2. It mentioned Laurent's case. I remember vaguely that Greek babies do not have their own names before one-year old, and they all use "Baby" as their name before the first birthday party. If that is the case, then will Greek-acquiring children all have at least partically similar articulatory filter just because they all have the same name? From: nagar@ling.ohio-state.edu The main point of the article is to show that there are individual differences in the phonological development other than those that can be explained with the help of biological and genetic differences among children. The other facts that stand out in the article (for me) are the claims that the linguistic practices of parents/caregivers does not influence the early phonological development of the child and that the name a child is called by has a bearing on her linguistic development. The questions that I would ask the author (which I think are rather naive) are: How universal is universal? How much data are we basing our claims on? What are the implications of your finding? i.e., if it is established that there is variability or no variability; how would that help our understanding of the way child language acquisition actually works? What is the influence of language internal and language external devices on language acquisition? In Other words, the kind of varability that you show; is it becasue of language internal factors or "universal" factors?