Institut de la Communication Parlée, CNRS UMR 5009,
ICP-Stendhal, BP 25, F-38040, Grenoble Cedex 9.
abry@icp.inpg.fr

abstract

MacNeilage gave in BBS 98 a thorough presentation of his "Frame/Content Theory of Evolution of Speech Production". In our comment, we supported globally his view that 7-months babbling cannot originate in Broca’s area. In addition to his arguments in favour of SMA (or a part of it), coming from cortical mapping and irritative lesions, we added a new piece of evidence, still unnoticed by its discoverers themselves: CVCV aphasia. We can now totalize an amount of about 60 patients, "brothers/sisters" of famous Broca's patient Tan, who actually repeated "tan, tan". Moreover such fluent "Tantan" or "Titi-Titi" global aphasics, some of them with lateral lesions, cannot support the proposal opposed to MacNeilage by three eminent students in motor control (Abbs, Jürgens and Lund), namely to lump into the lateral system both frame and content.

Concerning MacNeilage's view of babbling as a simple jaw based cyclicity, we are now able to demonstrate, by using articulatory models built from different subjects, that his "pure" labial frames are in fact idiosyncratic: labial, coronal or corono-labial, depending mainly on the anatomy of the baby (i.e. "why papa, mama…and tata" explained). Pushing a step farther, and given lips and tongue compliance, babbling variegation, could simply originate from biological noise in the motor signal generating jaw cyclicity.

A central question remains: what is the role of Broca? and when? Contrary to MacNeilage, we argue that a lot of speech behaviour in the first two years of life does not need to make a "loan to the intelligence" of Broca, as proponed by his BBS target paper. We will evaluate the benefit of delaying the appearance on the scene of such a morphosyntactic device as Broca’s, for the emergence of phonological universals, say the syllable related to first word prosody, and segment types related to articulator independence. This for ontogenetic developement.

Backwards, on the phylogenetic side, Rizzolatti made the same premature loan, when becoming acquainted of MacNeilage’s proposal. He argued on the base of the perception/action system, discovered by his team, called "mirror neurons", that neurons in the homologue of Broca’s area could match the "observation" and the "execution" of visual lipsmack communicative cyclicities. In fact, for the 3-day-old rhesus monkey, which demonstrates these affiliative lipsmacks, a "Broca" control is truly impossible, and Rizzolatti’s proposal is clearly a visuo-motorically biased misreading of MacNeilage’s target article, which insists both on the visual and audio nature of such a "precursor". Hence we need to go back to the non-lateral system (if not to SMA).

In fact Rizzolatti, like many other neuroscientists, seems to admit that non human primate vocalisations are not under cortical volitional control. This is obviously true when homologues of Broca’s and Wernicke’s areas (hence B-W) are lesioned. But not for the anterior cingulate gyrus (ACG, the head of the animal vocalisation system) and its neocortical extension SMA. We need to emphasize the endogenous, volitional control of this pathway, clearly defined as instrumental vocalization. For us this volitional aspect is the stem (and logically the necessary condition) of the vocal self monitoring system. This animal model of the vocal self has inspired a human model of auditory hallucinations in schizophrenia by Frith. One of its crucial feature is the use of B-W circuitry for the implementation of the corollary discharge hypothesis: this enables the recognition of one's own expected vocalizations.

We will claim that, for the moment, it seems impossible to get B-W circuitry involved farther in the evolution of language, without calling for an evolutionary Theory of Mind. This is precisely what is comprised into Frith’s and colleagues’ recent theoretical and experimental advances under the heading of a vocal alien monitoring system, found in the medial area BM 8. Whereas the elaboration of a Theory of Mind seems to us problematic with the action understanding system of "mirror neurons". It remains that, in our opinion, such a system is still in need of an adequate biocybernetic model, Jordan-Wolpert’s forward models, as the one used by Frith and colleagues in collaboration with Wolpert, being just able to implement the corollary discharge hypothesis, i.e. the initial vocal self monitoring system.

Our conclusion will be that, due to the focalization of research on the perisylvian system of language, the fractionation of the basic subsystems of speech has been waiting for too long. It's now time to look (i) at the basic system allowing the first steps of speech – babbling and its remnant in a very specific aphasia; and also (ii) at the two even more basic systems which allow people to talk to each other – the vocal self and alien monitoring systems. The co-evolution of these medial systems with the lateral ones will consequently appear less miraculous: (i) the emergence of the degrees of freedom for the carried articulators, i.e. the birth of their relative independence from the jaw, allowing the learning of language-specific segments on the basis of universal phonetic "places" and "manners" ; (ii) the connection with the medial cortex of the rich and specific systems for understanding the semantics of action, which is crucial for language learning.

 Conference site: http://www.infres.enst.fr/confs/evolang/