Home » Intelligent Design » The Cerebral Linguistic Toolbox That Blows The Mind

The Cerebral Linguistic Toolbox That Blows The Mind

“Depending on the type of grammar used in forming a given sentence, the brain will activate a certain set of regions to process it, like a carpenter digging through a toolbox to pick a group of tools to accomplish the various basic components that comprise a complex task” (1). This was the descriptive offered by one review on how it is that diverse regions of the human brain are recruited to tease out the meaning of sentences when we communicate with each other (1). Cutting edge research into brain function, using American Sign Language as a platform, has unpacked the detail of exactly how the brain achieves this split-second feat (1,2).

In sign language messages can be expressed in one of two ways. As with English, ‘signers’ can use ordered words to convey their message (eg: John gives his lunch to Mary). But they can also move their hands in a manner that specifically relays concepts and ideas- what linguists call inflection(2). In languages such as German and French inflections are easily identifiable as suffixes that can be tagged onto the ends of words to denote, amongst other things, the case or the gender of the word or the ‘role’ that a subject or object in a sentence plays in a given interaction (John giving lunch to Mary in the above example) (2). But sign language, notes Rochester University psychologist Aaron Newman offers “a unique opportunity to directly contrast these two means of marking grammatical roles within the same language” (2).

Newman employed functional Magnetic Resonance Imaging (fMRI) to zero in on the spatial-temporal brain activities that accompany both word order and inflection-based communication. What he uncovered was nothing short of remarkable. There exists a network of brain regions including the dorsolateral prefrontal cortex (DLPC), the superior and posterior temporal sulcus (STS), the caudate nucleus, the middle temporal gyrus (MTG), the angular gyrus (AG) and the left inferior frontal gyrus (IFG) that are operative during both the interpretation of word order and inflection processing (2). Importantly significant differences exist in the “relative weighting” of activation in these regions depending upon which of these two modes of message transmission is being called upon (2). The DLPC and the right hemisphere AG are more dominantly active when word order-critical sentences are put in front of us. In contrast the MTG and the posterior STS are more active during inflection processing (2). The overarching conclusion borne out by the results of this study is that “specific parts of the neurocognitive system recruited for grammatical processing are dependent on the type of information that must be processed” (2).

Over the years my interest in language and brain function has been fueled by my own exposure to cultures outside of those of my native England. I grew up speaking Portuguese, Spanish and to a lesser extent French. Unlike English, these and other Romantic languages display a requirement for word inflection in both verb endings and noun genders. Whereas English leans towards compound verb usage, Portuguese, Spanish and French show complex verb endings (e.g., The English phrase I shall come translates into Portuguese as Eu virei). When I traveled this month to the Brazilian Society of Biochemistry meeting in Foz de Iguassu in southern Brazil, I was relieved to find that I could slip almost effortlessly into both the written and spoken forms of Portuguese. It was a joy to find that, despite the odd non-conformity, my Portuguese had remained unadulterated over the years. Little did I know that I was employing cognitive functions that differed from those that I use in my more usual English setting.

On the flight back I settled down to read about the work of one Evelina Fedorenko who as an MIT psychologist has played an instrumental role in deciphering the functional hotspots of linguistic cognition (3). Her research has concentrated on mapping the ‘within language’ specificity (linguistic processing cognition) and ‘domain’ specificity (non-linguistic cognition) areas of the brain (3). Fedorenko and her close colleague Nancy Kanwisher have devised a localizer task approach for studying brain function (3). By asking individual subjects to perform cognitive tasks that place demands on localized regions of the brain (eg: contrasting pronounceable non-words like florp with real words like flop), they have been able to identify those regions that “engage in retrieving the meanings of individual lexical items and in combining these lexical-level meanings into larger meaning structures” (3).

Once back at home I had the chance to ask my father – a linguist by training – for his take on Newman’s and Fedorenko’s work. His principle observation was that sign language could only serve as a model for written words. Whereas sign language is sequential, spoken forms of language are multi-layered with sounds, grammar, vocabulary, intonation and gesture all acting together to achieve the conveyance of information. But what was plainly obvious to both of us was that through its sheer processing speed, the cerebral linguistic toolbox had no equivalent in anything that a carpenter might find on his workbench. Almost two decades ago brain biologist John Eccles noted that our linguistic capacity was pivotal in ensuring that we became the dominant species on our planet (4). The latest research is confirming Eccles’ assessment. And the brain architecture associated with language processing is turning out to be mind-bogglingly complex.

So, what of the evolution of language usage in humans? Perusing through the literature one finds a story that invariably begins with the need for some form of communication amongst early hominids. Having given up the safety of arboreal living in favor of an expansive conquest of terra firma, these hominids, we are told, would have relied on each other for information on the whereabouts of food, shelter and predatory dangers and may have been endowed with the simple descriptive function of language (4). “Cladistic branching with a great genetic change” accompanied the rise of Homo habilis, considered by many as the ‘initiator’ of spoken language (4). Great genetic advances gave us Homo erectus and finally Homo sapiens sapiens with his expressive, descriptive and argumentative capabilities (4). We are led to the idea that even this climactic achievement was accessible to the un-shepherded roving of natural selection (4).

Of course nothing in this story remotely addresses the question of how the interplay of diverse regions of the brain, such as that which we see above, became so firmly entrenched into the very fabric of how we communicate. The literature is silent about the details. As the twentieth century linguist Noam Chomsky argued language is “a skill that human beings are innately predisposed to acquire” (5). Today evolutionists are hard pressed to come up with an account for the origin of such an innate predisposition. At its core, the cerebral linguistic toolbox is a phenomenon that blows the mind and sabotages the evolutionist’s dream of a viable account for the origin of a vital part of our humanity.

References

1. Aaron Blank (2010) Sign Language Study Shows Multiple Brain Regions Wired for Language, See http://www.rochester.edu/news/show.php?id=3610

2. Newman AJ, Supalla T, Hauser P, Newport EL, & Bavelier D (2010) Dissociating neural subsystems for grammar by contrasting word order and inflection, Proceedings of the National Academy of Sciences of the United States of America, 107 (16), 7539-44 PMID: 20368422

3. Evelina Fedorenko, Functional localization in fMRI studies of language, See http://www.mit.edu/evelina9/www/funcloc.html

4. John Eccles (1991) Evolution Of the Brain, Creation Of The Self, Routledge Press, London, pp.95-96

5. Steve Blinkhorn (2003) Language Instinct, in The Science Book, ed. Peter Tallack, Weidenfeld And Nicolson Publishers, London, pp. 386-387

  • Delicious
  • Facebook
  • Reddit
  • StumbleUpon
  • Twitter
  • RSS Feed

7 Responses to The Cerebral Linguistic Toolbox That Blows The Mind

  1. Thank you for an interesting article. However I have two quibbles…

    Cutting edge research into brain function, using American Sign Language as a platform, has unpacked the detail of exactly how the brain achieves this split-second feat (1,2).

    Surely this is overstating their achievement? Uncovering some detail about a correlation regarding which brain regions are most active for which sorts of tasks is significant, but it seems a far cry from discovering “exactly how the brain achieves” the task.

    Whereas sign language is sequential, spoken forms of language are multi-layered with sounds, grammar, vocabulary, intonation and gesture all acting together to achieve the conveyance of information.

    I would have thought almost the opposite. Spoken language seems no less sequential than sign, and maybe more sequential. Sign of course has symbols, grammar, vocabulary, and its own forms of “gesture” and “intonation”. From wikipedia,

    Oral language is linear; only one sound can be made or received at a time. Sign language, on the other hand, is visual; hence a whole scene can be taken in at once. Information can be loaded into several channels and expressed simultaneously. As an illustration, in English one could utter the phrase, “I drove here”. To add information about the drive, one would have to make a longer phrase or even add a second, such as, “I drove here along a winding road,” or “I drove here. It was a nice drive.” However, in American Sign Language, information about the shape of the road or the pleasing nature of the drive can be conveyed simultaneously with the verb ‘drive’ by inflecting the motion of the hand, or by taking advantage of non-manual signals such as body posture and facial expression, at the same time that the verb ‘drive’ is being signed. Therefore, whereas in English the phrase “I drove here and it was very pleasant” is longer than “I drove here,” in American Sign Language the two may be the same length.

    However, your point about the evolution of language processing is unaffected by the above.

  2. I have always been fascinated by the brain, but skeptical about this wonderful organ to be the center of my consciousness. Having a dualist view of reality it is very easy for me to postulate that the brain is simply an organ that instantiate the intentions of my consciousness. Almost like a MODEM that modulate and demodulate messages from the physical world to the spiritual world and back. The “MODEM” activity should therefore correspond to all physical impulses as well as conscious thoughts / intentions or actions.

    All the reading on this subject that I have done has not convinced me that any of the observed brain activity is anything other than this “MODEM” acting as the channel to that try to ensure our successful action in the physical world. Because the brain is physical and subject to interference (due to any illness or other external action), it is possible to “mix or scramble” the messages to be interpreted by the soul / consciousness. I would say any brain condition can be explained as such. For example, inducing “happiness” by influencing the brain to signal the soul that it is the appropriate response to give etc. all falls within this “MODEM” paradigm of mine.

    My question is: Are there any evidence to support a purely physical interpretation of mind / soul / consciousness?

  3. mullerpr:

    My question is: Are there any evidence to support a purely physical interpretation of mind / soul / consciousness?

    This is an easy answer: no. All the evidence only points to how the brain processes the input and output of data to and from consciousness, and are absolutely compatible with, indeed IMO support, the model of the brain as an interface between consciousness and the physical reality.

    The “hard problem of consciousness” is a problem with no known solution. Consciousness remains an empyric reality which cannot be explained by an objective material model. While we know a lot more about specific details of how the brain works (but still it it only a bit of the real scenario), regarding the relationship between consciousness and matter we know nothing more than what was already known to the first man who discovered that a sensation was caused, in his consciousness, by an external event, or that he could in his consciousness decide actions which changed external events.

    In other words, that there is a flow of inputs and outputs between consciousness (subjective experience) and what we know as the external world, and that that flow is effective (creates correlations between the two realities) in both directions.

  4. I think it is important to stress the fact that you don’t need a materialistic view of mind to achieve any level of insight, through scientific investigation, in the the operation of the brain. In fact a dualist view of brain and mind creates the environment for scientific inquiry that has the possibility to find a physical mechanism for mind, should it exist.

    Since there seems to be no evidence for a materialistic mind it remains counter scientific to insist that the only hypothesis allowed will be a materialistic. Science has always been studying the effect of mind into the physical world.

    The phenomenon of algorithms being executed in/by machines is also clearly distinguished as the act of mind and not a new mind of its own. This include biological machines that constitute our physical bodies including our brains.

    With this said I have to ask another question:

    Are there any credibility in the notion of mind being an emergent property of brain? My experience is that it is nothing more than a materialist claim with no supporting mechanism or valid pathways described by such a mechanism.

  5. mullerpr:

    It seems we share the same ideas, maybe discussion would be more productive if some of the millions of fans of strong AI came here to contribute.

    Anyway, as you seem to want to provocate discussion about some fundamental points, I will follow you.

    So, emergent properties… Yes, I believe that consciousness has been labeled that in the AI field. The last I saw from Hofstadter, it seemed to emerge from loops. An old theory was parallel computing, but I believe that with all the quad-cores in circulation, none of which has come out to defend its rights as a sentient being, that position is a little bit obsolete.

    I find the concept of “emergence” at best elusive. I have just read the wikipedia entry for it, and I feel literally confused.

    Here are some pearls from that:

    Professor Jeffrey Goldstein in the School of Business at Adelphi University provides a current definition of emergence in the journal, Emergence (Goldstein 1999). For Goldstein, emergence can be defined as: “the arising of novel and coherent structures, patterns and properties during the process of self-organization in complex systems” (Corning 2002).
    Goldstein’s definition can be further elaborated to describe the qualities of this definition in more detail:
    “The common characteristics are: (1) radical novelty (features not previously observed in systems); (2) coherence or correlation (meaning integrated wholes that maintain themselves over some period of time); (3) A global or macro “level” (i.e. there is some property of “wholeness”); (4) it is the product of a dynamical process (it evolves); and (5) it is “ostensive” (it can be perceived). For good measure, Goldstein throws in supervenience — downward causation.” (Corning 2002)

    Is that a clear definition? Not for me. And think that darwinists always complain that the concept of CSI is ill defined!

    Is emergence to be find in structures, in patterns or in properties? What is “the process of self-organization”? (Prygogine, again?) What does “ostensive” exactly mean? What is maint here for “downward causation”?

    OK, let’s go on:

    The usage of the notion “emergence” may generally be subdivided into two perspectives, that of “weak emergence” and “strong emergence”. Weak emergence describes new properties arising in systems as a result of the interactions at an elemental level. Emergence, in this case, is merely part of the language, or model that is needed to describe a system’s behaviour.
    But if, on the other hand, systems can have qualities not directly traceable to the system’s components, but rather to how those components interact, and one is willing to accept that a system supervenes on its components, then it is difficult to account for an emergent property’s cause. These new qualities are irreducible to the system’s constituent parts (Laughlin 2005). The whole is greater than the sum of its parts. This view of emergence is called strong emergence. Some fields in which strong emergence is more widely used include etiology, epistemology and ontology.

    So, we have weak emergence and strong emergence. Weak things, obviously, are not the most interesting. But does strong emergence really exist?

    Regarding strong emergence, Mark A. Bedau observes:
    “Although strong emergence is logically possible, it is uncomfortably like magic. How does an irreducible but supervenient downward causal power arise, since by definition it cannot be due to the aggregation of the micro-level potentialities? Such causal powers would be quite unlike anything within our scientific ken. This not only indicates how they will discomfort reasonable forms of materialism. Their mysteriousness will only heighten the traditional worry that emergence entails illegitimately getting something from nothing.”(Bedau 1997)

    Ehm… Perhaps it’s better to look at practical examples:

    Non-living, physical systems:

    Some examples include:
    Classical mechanics: The laws of classical mechanics can be said to emerge as a limiting case from the rules of quantum mechanics applied to large enough masses. This may be puzzling, because quantum mechanics is generally thought of as more complicated than classical mechanics.
    Colour: Elementary particles do not absorb or emit specific wavelengths of light and thus have no colour; it is only when they are arranged in atoms that they absorb or emit specific wavelengths of light and can thus be said to have a colour.

    Patterned ground: the distinct, and often symmetrical geometric shapes formed by ground material in periglacial regions.
    Statistical mechanics …
    Weather
    Temperature is sometimes used as an example of an emergent macroscopic behaviour. …
    Convection in a fluid or gas is another example of emergent macroscopic behaviour that makes sense only when considering differentials of temperature.

    In some theories of particle physics, even such basic structures as mass, space, and time are viewed as emergent phenomena, arising from more fundamental concepts such as the Higgs boson or strings. In some interpretations of quantum mechanics, the perception of a deterministic reality, in which all objects have a definite position, momentum, and so forth, is actually an emergent phenomenon, with the true state of matter being described instead by a wavefunction which need not have a single position or momentum. Most of the laws of physics themselves as we experience them today appear to have emerged during the course of time making emergence the most fundamental principle in the universe and raising the question of what might be the most fundamental law of physics from which all others emerged. Chemistry can in turn be viewed as an emergent property of the laws of physics. Biology (including biological evolution) can be viewed as an emergent property of the laws of chemistry. Finally, psychology could at least theoretically be understood as an emergent property of neurobiological laws.

    Well, let’s see: what is left out of the list of emergent properties? Maybe politics? No, I am sure that too emerges from something.

    And now the best:

    Living, biological systems
    [edit]Emergence and evolution
    Life is a major source of complexity, and evolution is the major principle or driving force behind life. In this view, evolution is the main reason for the growth of complexity in the natural world. If we speak of the emergence of complex living beings and life-forms, we refer therefore to processes of sudden changes in evolution.

    Well, now my point of view. Consciousness is an empiric reality. We are conscious, and we know we are. Conciousness is experience directly in one’s consciousness, and inferred by analogy in others’ consciousness. That point I have made many times here.

    Material objects apparently are not conscious. If consciousness has to be explained form a purely material model, we need evidence of two kinds:

    a) the model should be able to explain why some properties or patterns or structures of material objects should be connected to the existence of a conscious experience.

    b) experimental research should point to examples where the objective configuration of material objects produces a subjective conscious experience.

    I have never seen anything even remotely near to accomplishing either a) or b).

    So, to sum up:

    1)I would say that I have really no idea of what is meant by emergence, and even if I could get some better understanding, I can’t see how it could be of any use in explaining consciousness.

    2) There is no logical reason why any specified structure or pattern should be able to cause subjective experiences.

    3) There is no empirical evidence of that.

  6. gpuccio,

    Thank you for the highlights from Wikipedia, but it is clear that your criticism is the most valuable view about the subject.

    It seems as if the proponents of emergence is not going to enlighten us. Maybe it is because they are of the same mind as the Wikipedia contributors of the subject.

    I am not sure if emergence explains everything or if it explains nothing at all. Impotent concept comes to mind. Contrast that against the excitement and usefulness of a pursuit of investigating our own and the acts of any other consciousness. (…a scientific take on “love thy neighbor and love thy God.”)

  7. Just a short comment which is specially pertinent in the ID context.

    The cerebral linguistic toolbox is certainly a prominent feature of the higher organization of the human brain. While we are still completely ignorant of which are the characteristics of human brain which make it so capable of higher functions which can be found only in primitive form in other animal forms, and while the genetic origin and control of that higher organization remains a complete mystery, the ID theory gives us at least one objective mark of the superio functions of the human brain: it can generate CSI. Indeed, it alone can generate CSI in the known universe, with the only important exception of the designer(s) of biological information.

    This is very important, because it means that the human brain, acting as an interface for human consiousness, can do what any computer, however complex, cannot do: it can produce non algorithmic CSI. And language is probably the first form of CSI in human history.

    So, the fact that the processing of language at conscious level requires such complex and integrated working at the brain level is really interesting. Any new information about these issues will help describe the physycal processing which takes place in such higher functions, and in time elucidate better the role of conscious representations, and of their interaction with physical algorithms and activities.

Leave a Reply