Roger,
I was taught in graduate school that the ability to make and hear phomenes of all language is inborn, but as you learn a language, you lose the ones you don't need, and after about age six, it's hard (or impossible?) to get them back. I also heard that there was some language school in Boston that goes at that problem head on and gets back those phonemes for you. But this is the sort of stuff you probably know more about than I do so I won't say more. I am trying to sing German and Italian right now and both require a trilled "r". I simply cannot do it. The closest I can manage is singing the German r as a "d", which is close enough for gummint work, but very unsatisfying. I also can't blow a proper raspberry. Never have been. If you run into somebody in Santa Fe that recovers lost phonemes for old people, let me know. Nick From: friam-boun...@redfish.com [mailto:friam-boun...@redfish.com] On Behalf Of Roger Critchlow Sent: Monday, February 13, 2012 12:04 PM To: The Friday Morning Applied Complexity Coffee Group Subject: Re: [FRIAM] FW: Re: Friam Digest, Vol 104, Issue 9 I was torturing myself yesterday listening to Spanish instructional material in my car. The tapes spent a lot of time presenting minimal contrasts between vowels in different contexts, between consonants, or between alternative stresses. I can hear some of the contrasts quite clearly, I can hear others if I listen carefully, but there are some where my auditory system just goes "what?". I'm hoping that reviewing the tapes with the text in front of me will help, or that I was just getting tired. I suspect that someone with a differently accented american English upbringing might have trouble with a different subset of the contrasts. I think there's a parallel between all of this and http://web.mit.edu/newsoffice/2012/error-correcting-codes-0210.html. The decoding of linguistic communication is an omnivorous error correcting algorithm which operates in situations with unpredictable noise. It will use anything associated with the communication act to aid a decoding, but it usually gets by with the simplest and most straightforward coding. And it decodes all the alternate code words, as it were, in parallel. So we're happy to decode scrambled words, as long as the word length and letter distributions are correct and the scrambled words make a sensible utterance. Why not? We're perceiving word lengths and letter distributions all the time when we read. Similarly, but harder to demonstrate in a typed email, my calligraphy teacher once showed us that a line of text can be read from only the tops of the letters. The bottoms aren't nearly as informative. -- rec -- On Mon, Feb 13, 2012 at 10:41 AM, Eric Smith <desm...@santafe.edu> wrote: Eric, hi, There is a (to me) fun similarity between this sequence you have illustrated, and some pathologies and treatments in rapid auditory processing, in which the workers I know are April Benasich and Paula Tallal (this, from a few years ago). The stopped phonemes, particularly the voiced, non-plosives (in English, b, d, g) differ by transients that are on the order of <<100ms in duration. A subset of children are born with deficits that render them unable to resolve transients this short in the primary auditory cortex (I don't know or remember what the right reference is to the brain region or layer). As a result, they can't hear difference between /bag/ and /gag/, and they grow up with severe dyslexia through having never had needed queues to the structure of the language. The human vocal tract can't alter these stops, because they are mechanically determined, so there is nothing the caregiver can do. What the remedy is, is electronically altered speech, which can slow the transient down while retaining the structure of the formant timeseries that distinguishes the segments, so that it spans more than 100ms, while leaving the rest of the word length and overall speech prosody the same. The child is played a few hours of this altered speech per day, through the early formative few years (I forget, perhaps pre-9-month to 3 years). The child develops near-normal language and reading skills, and the altered-speech therapy can be phased out. After that, the child functions essentially normally. In sentence context, he believes he is hearing the distinction between /bad/ and /dad/, and processes sentences normally, though in fact this rapid-auditory deficit is congenital and never develops normal function. However, there is so much redundancy built into the combination of phonology, lexicon, morphology, and syntax, that it can serve many functions over the course of lifetime language use, from child learning, to error correction in normal speech, as in noisy environments, or correcting for speaker accent variants. Our colleague Morten Christiansen even has evidence that phonotactic queues provide significant evidence to distinguish word categories such as noun and verb in English, which child language-learners are using, along with morphological queues. Somehow, all this works, because there is a complicated mix of top-down conditioning of fine-scale perceptual processing, updated by bottom-up updating from the actual rapid inputs. I have assumed that things like this are also going on for written language, and I have also assumed (though in this case without defense from neuroscience), that this is an expression of a general feature of the multi-level character of neural processing, which can also be seen in the function of systems such as reflex arcs, which, e.g., require different responses for reaching to touch, versus to grasp, even though they use the same afferent and efferent systems, and are conditioned by a reflex arc rather than going all the way back up to tactile/motor cortex interactions. I don't know whether this addresses the kind of explanations you think would be relevant or appropriate to phenomena such as this. All best, Eric On Feb 13, 2012, at 7:28 AM, ERIC P. CHARLES wrote: Benny, The first quote is the only one I think is from me. To clarify: When you say that the phenomenon is "reading gibberish", then it seems like it might be a skill. However, if you phrase it as a failure to distinguish gibberish from properly written words, or as mistaking gibberish for properly written words, then it seems less impressive. Let's set up a hypothetical longitudinal study... At age two you show your child two words "amazing" and "azanmig" and verbally ask them which one is the word "amazing". The child guesses at a 50/50 rate, and from that we conclude that the child in unskilled. Later, at age 4, the same test is given, and carefully sounding out the words, the child identifies the correct word every time. This, we conclude, demonstrates growing reading skills. Next, lets say, thirty years later, we send the child an email, in which, in the middle of a sentence, there is a claim about the azanmig powers of the human mind. Our (adult) child reads the sentence without even noticing that there is a mistake. Why would the failure to perform at the level of a 4-year-old be suddenly taken as a sign of advanced sophistication? Surely there is some way to argue that the final incident indicates an advanced skill of some sort, but our initial bias should be to assume that failures to discriminate indicates a lack of skill. Is that more clear? Eric On Sun, Feb 12, 2012 06:15 PM, Benny Lichtner <benjamin_licht...@brown.edu> wrote: Hi, all. Just want to respond to some of these points as I am very interested in meaning making. It [the ability to read "gibberish"] does not demonstrate mysterious skill, it demonstrates a (perhaps mysterious) lack of skill. The real mystery, if there is one, is why a person so well trained in reading would be fooled by such a simple manipulation. I do not see where the "lack of skill" is. Could you elaborate? What skill is it that we lack because of our ability to read "gibberish"? Is it the skill of identifying what is "gibberish" and what is not? The problem with "gibberish" is that you always have to put scare-quotes around it. I have to ask why even an intelligent person can make some kind of sense out of ["]Gibberish.["] Something about our brains seems to Beg for Sense when there is none. I suspect that were our brains not to "beg for sense," we would never be able to read any non-"gibberish." There are no hard rules for generating meaning, just some customs we use, and are used to, and break much of the time. We prefer magical explanations because they do not require any effort. Is it that we "prefer" magical explanations, or just that they are the first to come to mind, and are perfectly satisfactory explanations at that, so why look for others? If the audience were simultaneously presented with a magical and a non-magical explanation, I am not sure which would be preferred. It might have nothing to do with magicality. Maybe more to do with creativity, or realism, depending on who you are. All Physicists are apparently insane because they intentionally look for difficult explanations. The explanations that physicists look for are difficult by coincidence, I think. They look for explanations that fit their assumptions about knowledge and causality, and those explanations happen to be difficult, both to invent/discover and to understand. Apparently Cryptic equals Important for most people. I wonder if this has something to do with the reward that comes from understanding something challenging. Cryptic (wo)man-made things may also seem important because it is mysterious why a (wo)man would spend time creating something cryptic. --Benny Lichtner Eric Charles Professional Student and Assistant Professor of Psychology Penn State University Altoona, PA 16601 ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org ============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
