Talk:Dunstan Baby Language

From Wikipedia, the free encyclopedia

Hi. I have removed the references to Brown University, since the claims on Dunstan's web page concerning the involvement of this institution are doubtful. Here's why:

  • There is no "Language Research Center" at Brown University. There is a "Metcalf Language Research Lab", but the name Priscilla Dunstan does not occur anywhere on the institutional web pages, nor anywhere else on Brown's website. The research projects described on the language lab's web page do not make any reference to Dunstan, and there is not a single publication on this topic.
  • The claims seem to be deliberately worded ambiguously: The first mention is on the "Research" page [[1]], where it is claimed that what is reported is "Research conducted by Leading Edge research, Sydney Australia with the involvement of Brown University, Rhode Island." What does "with the involvement" mean? Additionally, at the top of the page it says "Independent research conducted in Sydney and Chicago with over 400 mothers in 2006 found some remarkable results" -- but Brown University is not in Chicago. The second mention is on tje "Meet Priscilla" page [[2]], where there is mention of "years of independent studies - including those guided by Brown University’s leading infant research centre" -- this could actually be read to mean "unrelated research". In my opinion, these ambiguous wordings are deliberate. They are meant to give the impression that Brown University is involved without actually saying so.

If anyone can turn up hard evidence of a link between Dunstan and Brown University, then this can be mentioned in the article, but until then we should be careful not to spread unfounded PR claims. MadProfessor42 08:16, 7 December 2006 (UTC)

I suppose it's possible someone will find hard evidence of a link between Dunstan and Brown, but I kind of doubt it given this find: "Journalist Jane Cadzow did an excellent job of unearthing disparities in their claim of 'independent studies'. [. . . . ] Brown University disputes this stating in the article 'Brown University faculty did not conduct this, or any, research on the Dunstan Baby Language system'." [3] KingTor (talk) 05:10, 17 December 2007 (UTC)

[edit] Photo of Priscilla Dunstan

I feel that the photo of Ms. Dunstan should be moved from the left side of the page to the right, because it's interfering with the alignment of the 'See Also' section. —The preceding unsigned comment was added by 68.51.173.99 (talk) 19:28, 9 February 2007 (UTC).

[edit] scientific

The article currently says "Dunstan's claims, in their current form, do not constitute a scientific theory, and ..."

Is that so? The basic idea -- that different kinds of discomfort cause babies to make different sounds, and that people can learn those sounds and use them to figure out which kind of discomfort the baby is in -- seems to me that could be tested in a double-blind test, so it is at least "falsifiable" (falsifiability) and therefore a "scientific theory".

--nope, that's what we call an hypothesis. 69.7.77.20 (talk) 16:34, 29 May 2008 (UTC)

Whether or not anyone actually has tested it, and whether or not the test falsified the idea (or, in fact, whether the idea is true or false), is irrelevant to deciding whether or not something is a scientific theory. --68.0.124.33 (talk) 06:47, 4 January 2008 (UTC)

You suggest a double-blind test would be possible, but a double-blind test requires control and experiment groups. I highly doubt that would be possible in the case of babies, since there is no (ethical) way of assigning babies to groups that would experience pain of a certain type. Sure, we could starve one group and see if those babies say "Neh" more often than the control group, but I doubt you'd find any parents willing to subject their infant to that. So maybe it's technically falsifiable, but I don't think she ever expects her "theory" to be subjected to that kind of rigor.
I've actually seen the DVD. She has riddled her system with so many caveats and exceptions that basically it boils down to this: "If your baby's making noise, try to comfort it with food, burping, changing, and/or sleeping." It's not falsifiable because it's essentially a tautology. She's tailored it to be always true. --W0lfie (talk) 23:34, 4 January 2008 (UTC)

Yes, a double-blind test requires control and experiment groups. And I agree that confirmation bias makes this system difficult to test.

I agree that experimenting on babies is probably unethical, but perhaps there is another way. Off the top of my head (original research warning), here's one possible test (perhaps there is a better one): We videotape babies. I think we could find parents willing to subject their infant to videotaping. No one deliberately subjects the infant to pain, but (as always happens with every baby) eventually the baby starts crying, and the parents (eventually) figure out why, fix the problem, and the baby stops crying. The parent writes down (or tells the camera) what the problem was.

We end up with pile of video tapes, and reduce them to one snippet for each time each baby cries -- a little snippet that lasts a few minutes (?) and ends with the child starting to wail, and piece the snippets together (in some order? randomly? how to avoid unconscious bias?) into a movie. Much later, we get 2 groups of adults -- a control group and an experiment group. Each adult in the experiment watches the movie and for each snippet guesses (a) which sound the baby is making, and (b) what the problem is.

If the theory is correct (that the language is "universal", applying to every baby), then the experimental group of adults who have been given "language training" will correctly guess the problem significantly more often than the control group. Also, if the theory is correct, for any particular snippet, most adults will agree on "which sound is this baby making" during that snippet. Also, if this theory is correct, there will be a strong correlation between the guessed "what sound the baby is making" and the actual problem. See any fatal flaws that would make this "scientific" experimental protocol biased or unethical? --68.0.124.33 (talk) 18:09, 10 January 2008 (UTC)

I guess it's no worse than several other scientific theories that are tested under a presumption of correctness. But I would argue that if it's truly a universal language, it wouldn't need teaching. Look at Eckman's research into universal expression of emotions. That's universal. People just know what a smile means. Well, this is getting off-topic. I apologize.
Back to your suggestion, I suppose it would be a valid experiment to verify whether or not the language training helps parents solve their baby's woes. Which makes for a more externally valid experiment at the expense of its internal validity. Such trade offs are typical in experiments where ethical concerns prohibit a completely valid study. And frankly, I would settle for even an externally valid experiment such as the one you suggest. So I guess if the "theory" in question is the teachability (is that a word?) of the language and its real-world usefulness (rather than the definitions of the "words" in baby talk), then I would agree with you. It should count as a scientific theory. --W0lfie (talk) 02:53, 12 January 2008 (UTC)
Almost forgot. The sentence as worded says "Dunstan's claims" aren't a theory. So I'd still agree with that statement, since her claims include a whole slew of things that aren't wholly falsifiable, due to their tautological construction. God, I sound pedantic. Sorry. --W0lfie (talk) 02:57, 12 January 2008 (UTC)