Comentários de William Powell enviados em 07/07/97 acerca do Sitio da Mente

I enjoyed your book (sitio da Mente) very much desoite the fact that it took me more than 40 hours to read it. Because I read Portuguese slowly, I had time to think as I read it. It stimulated many thoughts and this is why I enjoyed it so much. I know a German edition may be producel so I will be very candid for you and tell where I think it can be improved. I will start the simple errors I noticed:

1 figure 15 on page 72: The ponds (Ponte) is not correctly labeled.

2 Page 97: The chapter title is missing a space. I could not find the word ETELAS in my Portuguese/ English dictionary. Finally I realized it should be E TELAS.

3 Page 108 figure caption: I may have misunderstood the Portuguese, but it seems to me that you are stating in this figure caption that the boxes are only related to visual processing. If that is a correct reading of the caption, then it is wrong as I am sure you already know. I believe D.J. Felleman first drew this figure. It is somewhat arbitrary where one should stop considering these different neural regions to be concerned exclusively with vision, but no one would consider vision to be even a primary function of the top two boxes (ER and HC). (The HippoCampus and the EntoRhinal cortex are associated with memory, much more than vision.) In their paper (Science 225, p 419) Van Essen, Anderson and Felleman state: "Twenty five areas are primarily visual in function; the remaining seven are also implicated in other functions ..."

4 Page 111, first sentence of new paragraph: Again I am sure that as a doctor you know that the brain does not regulate the level of oxygen, and that the stimulus to breath is an elevation of the carbon dioxide level in the blood. You may not know that this fact can have tragic consequences. Two men died at APL, where I worked, because the brain does not control the oxygen level. - They were working on a satellite inside a large steel vacuum test "bell jar" that they had cranked up only enough to crawl under. They had also left the dry nitrogen supply line open. Slowly the oxygen level inside the bell jar decreased. They continued to breath out carbon dioxide and as this is what is controlled, they remained undisturbed by the decreasing oxygen content of the "air" inside the bell jar. They were not found in time to revive them. They could have left at any time but had no desire to do so as their brains kept their blood levels of carbon dioxide in the controlled range until they died of oxygen starvation.

5 Page 125, first sentence: On the prior page, first new paragraph, you have equated "mente" and "consciencia," thus it seems to me that the first sentence of page 125 is a direct contradiction of the second sentence in the second paragraph of page 109. As I have read the entire book, I understand that you are making a strong distinction between mind and brain, but still you should not contradict yourself. I again note that this "contradiction" may also be just my lack of full understanding of Portuguese.

6 Page 175, line five from the bottom: I wish it were true, that competence in language were innate as then I could read your book with ease. What is innate is the ability to acquire language (s). What Chompsky calls the "language acquisition device", or LAD is innate, not the competence in any particular language. It might also be noted, that not all experts agree that the LAD is innate. Hillary Putman, for one, has long argued that there is no special LAD. According to this minority view, language is learned with general learning facilities of the brain, nothing special or innate.

7 Page 476: Footnote 18 appears twice. This is not a simple typo, but an inversion of sequence. That is, the first 18 goes with the text reference 19 on page 185 and all remaining must be renumbered in the text as well.

This is the all the "errors" I noticed. There are several places where I think your text is misleading but not actually wrong. Again it may only be misleading to a foreign reader, but I will list the places that concerned me so you can evaluate them again:

1 Page 32: You over emphasize the importance of brain size and do not emphasize the importance of brain organization enough. I note that Neanderthals had larger brains than our ancestors. Also elephants have much larger brains. Anatole France’s brain weighed only 900 grams, less than most apes.

2 Page 81: To state that the retinal information goes to the occiptal lobe ignores the fact that at least 95% of the data available in retinal cell activity is never sent to the occiptal lobes (or even to the LGN). The extent of data compression processing, (retinal extraction of the more relevant part of the information) that occurs in the retina can be estimated by comparing the number of optic nerves to the number of light sensitive cells in the retina.

File compression routines in your computer preserve all the data, but some is discarded in the eyes. As you can recover the scene viewed with surprising accuracy from the small amount of compressed data sent over the optic nerves, it is true in some sense, that the data has been sent to the brain, but certain illusions exist related to "filling in" that prove that some data has been discarded in the eyes (unlike computer file compression).

3 Page 109, the paragraph continued on to page 110: What you state is true of short term memory, STM, but when you only call it an example of memory, I think most people will understand "memory" as long term memory, LTM. As you know well, LTM is associated more with the changes in synaptic connection strengths (at least if there is any truth to the currently accepted view). I don’t think you should allow readers to assume that your triad applies to LTM and this is what I suspect most will assume from the text as it is now.

4 Page 342 & footnote 1 page 489: Perhaps the word "amplitude" has more meanings in Portuguese than it does in English, but if it only means size, then it is not the amplitude of action potentials that increases with LTP, but only the synaptic connection strengths. In English the word "potential" has two meanings. The meaning related to amplitude should not be confused with the meaning of the word "potentiation", as used in LTP. LTP proceeds mainly by post synaptic changes, but there are some presynaptic changes with memory formation as well. (I have references if you are interested as the point is relevant to the "back propagation of error" as a means of training artificial neural networks.) You could accurately say that the amplitude of the synaptic transfer is increased by LTP - i.e. the quantity of neurotransmitters and/or their residence time in the synaptic junction and/or the number of channels is increased by LTP, but not that the action potential is increased by LTP.

5 Page 464 footnote 2: (I disagree with this footnote almost entirely, but that does not make it wrong.) I think it may be slightly misleading when discussing "alfabeto de tipos" and later "type identity." With a degree in philosophy and years of experience and interest, I am sure you understand the "type identity" concept more fully than I do, but my impression gained from this footnote is that you want at least the structure of the "bar codes" to be constrained by some innate universals where as I understand "type identity" not to require anything with regard to the instantiations, not even their abstract structures. That is, "type identity" would in no way constrain the "bar codes" or any other instantiation of information. I understand "type identity" to only constrain the functional roles, not any datail of the implementation (s).

Perhaps my confused impression is only do to my lack of experience with Portuguese, but perhaps you can your point more clearly so my confusion about exactly what you mean is reduced.

Notes:

1 I would first like to note that we share a minority view. I too believe that much of the mind is in the temporal realm, more than the physical, and that this helps with the mind/body problem (free will etc.). I am not willing, however, to go as far as you do with the idea that the mind’s "bar codes" need not be associated with specific neural groups. This separation of mind from specific neurons is a theme that appears many places in your book, but most clearly and explicitly in several sentences on page 98 and the last paragraph of page 350, where you briefly consider what I will call the "Penfield argument" and reject it. I find the Penfield argument more persuasive than you do, but it is not the strongest argument against totally disconnecting the "bar codes" from specific neural tissues. I have several arguments which force me to disagree with you on this point. One I will call the "RSVP argument" (RSVP is for Rapid Serial Visual Presentation, not Respondez S’il Vous Plait). This argument vaguely resembles the "100 step argument" against LISP type serial processing. The RSVP argument considers (Fact 1) the maximum action potential discharge rates and (Fact 2) that humans can read more rapidly than 30 word per second (See Vision Research, vol. 32 #5, p895-902 and note that the frame rate of the display, rather than subject’s reading rate, established the published limit of 30 word/second!) These two facts combine to set an upper limit on the length of the "bar code." One can argue exactly what the maximum "bar code" length in bits (or number of action potentials) can be, but these two facts force it to be approximately 30 neural spikes or less even if one assumes each word requires only one bar code to capture all its meanings, limitations and competence in use. Thus the human mind could think no more than approximately 2 exponent 30 distinct thoughts. Furthermore, this large number is greatly reduced as most of these bar codes could not be used, unless one is willing to make the very unreasonable assumption that the brain has a precise time clock. Not only would all codes with leading or trailing nulls (zeros or absences of neural pulses) be unusable but all those codes with more than four or five repeats of either 1s or 0s would be confused. E.g. 100000011001 would be confused with 10000011001 and likewise 1111111001100110011 with 111111001100110011. etc.

I have given only a weak version of the RSVP argument because I have hard data only for it, but if one considers that each letter of the alphabet requires a code and a skilled typist can generate these 26+ different codes at a very high rate this fact may set a lower upper limit on the code length. Another possible way to set the limit lower may be the fact that ordinary TV occasionally makes an RSVP sequence of scenes. I am sure I recognize dozens of objects in each when I have watched one of these TV RSVPs. I am not sure of the frame rates used, but suspect a TV type RSVP, of objects in color, would force a lower bar code length than the hard RSVP reading rate data I know about. (I only knew about the cited Vision Research paper because I knew the authors at JHU.)

Because I keep the bar codes of my version of the mind tied to specific neural tissue, I am not a victim of the RSVP argument as you are. That is I can use the same bar code for as many different concepts as I am willing to postulate distinct neural group assembles. Since each neuron could be (and in fact is) used in many different neural groups, and there are a huge number of neurons, my model of the mind will never run out of bar codes no matter how many distinct thoughts I can have. I could even (but do not) postulate only one bar code of only a few bits!

2 Another theme that runs through your book is that the mind is based on a chaotic brain. If you mean only that attractors exist, I certainly agree. Any oscillator can be considered to have at least one attractor in the frequency dimension of phase space. I believe that the mutual interaction that I described in the reprint I gave you, leads to the so called 40 hz oscillations and I tend to accept the popular idea that "the binding problem" is at least partially resolved via this correlation of physically separated action potentials. I do, however, have some very unique (strange?) ideas about why binding of features is required and how it is used.

If, however, you mean that the mind is very sensitive to "initial conditions" or "inputs," as I think you do from reading your book, then I must disagree here also. I think you hold this view in an effort to escape from determinism and open the possibility of free will. I of course support this goal but fear your aproach can only result in chance, not the free will I want. In addition to being a road that leads to chance, not free will, this stronger version of a chaotic mind, would seem to me to be impossible to develop by evolutionary processes. Not only does ever creature that evolved tend to "status" (medical term, that I assume is as much a part of a Brazilian doctor’s background as it is for a US trained doctor), but every organ of every creature has evolved to seek stability, not chaos. If I am to accept some process other that evolution as the source of the mind, I might as well become a Cartesian Dualist.

3 No one likes to be put in a box with a label, but if you will forgive me this error, I find your philosophical views (as I surmise them from your book) to be a most interesting and a quite rare combination. You seem to be a "properties dualist", yet you believe in universal innate codes (or at least in universal innate structures with only adjustable parameters to vary instantiation codes). I can’t help wondering why one who believes that these universal codes (or structures) exist does not also believe that, at least in principle, the mind should be reducible to these universals? If I have most unfairly forced you into inappropriate boxes, perhaps you may wish to clarify this appatent conflict in the German edition.

4 There are many more things in your book of great interest to me. It is truly one of the most interesting books I have read. But I must tell you that I think you can and should rewrite chapter 18 (language) for the German edition as in my view you know much more than you state there. (It is always nice to claim new material in the introduction of each new edition.) Language research has lead the cognitive revolution and you note this only in other parts of the book. (In chapter 18 there is very little that James could not have written 100 years ago!) I dont’t think you should go into the details of Xbar theory, but at least you, as a doctor, should describe the difference between Broca’s and Wernicke’s symptoms. I.e. say something like: "a broca’s victim has difficulty in producing speech but good comprehension and note Broca’s area is near the motor cortex; whereas a Wernicke’s patient talks a lot, but has little content and poor understanding of speech" Note that Wernicke’s area is in the temporal lobe near the auditory cortex and the cortex that vision uses to determine the "what" as opposed to the "where" of an object -i.e. near other cortex also used to give "meaning." After reading you excellent description of depression (pages 266 through 269) I was surprised by your shallow treatment of aphasia. If it were me, I would also mention "conduction aphasia" even though this term is not used as much as the other two. I like to mention it because the symptoms are exactly what you would expect if the "wires" from Wernicke’s area to Broca’s area were broken -i.e. the patient talks like a Wernicke’s area victim with little meaning in his fluent speech, but unlike a Wernicke’s area victim, he understands speech well. (Nothing wrong in the temporal lobe, it is just that the ideas can’t get to Broca’s area to be formatted into "surface structure" sentences and speech commands because the "wires" are broken.) This dependency of ideas on specific neurological areas is another of the many reasons I can’t go a long with the concept that the mind "floats" in time independent of specific tissue as already discussed in part in Note 1 above. I call this the "Aphasia argument" against bar codes independent of specific neural groups.

5 There are many other things I hope to discuss with you in person some day. I am extremely happy to make contact with some one like you who knows a great deal about how the mind and brain function and some of the related philosophical questions. I have restricted myself in this already too long letter to items that may help you in the production of a German edition. I trust you will welcome the critical comments I have made, but again wish to tell you that I found your book very interesting and stimulating.