Developmental Heterochrony
This is an extension of the discussion in "Learning Academic Language Identities: Multiple Timescales in the Social Ecology of Education" (UC Berkeley, March 2000) of a new approach to conceptualizing developmental change.
The specific argument is that (1) individuals develop in the normal context of age-diverse communities, and (2) developing individuals, by interacting with an increasingly wide range of individuals of different ages, come to incorporate more and more of the full behavioral repertory of the community as a whole. Moreover, development takes place on multiple timescales, each generated by developmental processes themselves, so that a simple clocktime model of linear developmental progression is inadequate, since processes on very short (fast) timescales to very long (slow) timescales are interdependent.
I propose that a more adequate way to conceptualize development is to imagine that each developing individual is characterized by a distribution of probabilities or frequencies of displays of elements of its total behavioral repertory, for particular situations, but especially as integrated across situations that evoke different parts of the repertory. For developmental analysis, we focus on some particular kind of behavior (e.g. language use, motor habits), and identify the most frequent or typical behavior for each age-cohort in a community. These then become, in sequence, proxies for developmental time. At any given chronological age, an individual is characterized by a probability or frequency distribution function across the behavior-chronology range.
Because behaviors tend to remain in the repertory even when later behaviors become more typical in a given situation, and can be evoked under some circumstances, there is always a lagging edge tail to the distribution function. In the view of behavior being used, all behaviors are relational, to other participants (animate or inanimate) and to the situation. So we might expect that more juvenile behaviors surface from the repertory in some situations where the individual is interacting with younger individuals (though this setting may also evoke precociously 'older' behaviors as well).
Because as the individual ages, s/he accumulates interactive experience with more and more older members of the community, and because successfully learning to negotiate these interactions involves, to increasing degree, the ability to model (at least for comprehension or for smoothly coordinated joint action) the elder's ways of participating and behaving (this includes observations of elder-elder interactions as well as participation in self-elder interactions), there is also an anticipatory (or 'precocious') leading edge to the distribution function.
The present hypothesis is that development consists in a change in the overall distribution function so that it comes to more and more model the collective distribution function of the community as a whole. But equally significant is the basic picture of developmental "stages" as modeled distributionally for the individual, with both 'regressive' and 'precocious' behaviors at all stages, particularly visible in age-heterogeneous settings. The model as a whole is antagonistic to a strict "stage" perspective, because development is running simultaneously at many different rates, by many different internal clocks, so that even a discontinuity on one timescale is underlain by relative continuity at the scale below (averaging over rapid changes) and the scale above (relative constancy).
A rough picture of developmental heterochrony (simultaneous distribution across the developmental range of age-typical behaviors or characteristics) is given in the diagram below:
Each curve in the upper graph is an instance of the developmental distribution function for an individual at a particular age. Each curve further to the left represents the same individual at a later age. The vertical axis is frequency or probability of display of a behavior of a particular kind (in general a phenotypic character) in a particular setting, or more generally averaged across all possible settings; the horizontal axis represents a proxy for chronological time (NOT the age of the individual) formed by taking an index of the behavioral practice of the kind being graphed in the distribution which is most frequent or typical in the community for individuals of the corresponding age. Development of the individual corresponds to the sequence of distribution functions, the change in location of the peak and the change in shape, with age. For an individual who was at every point in development exactly typical of all age-peers (an idealization), the peak of the successive distributions would always fall at the exact proxy for his/her age.
For a real individual, the peak at any given age corresponds to his/her own most frequent or probable behavior at that age; the trailing edge corresponds to behaviors typical of younger ages (for the community as a whole, not necessarily this individual's own behaviors at that age, which would be represented in a previous function to the left); and the leading edge corresponds to anticipatory or precocious behaviors that are more typical of other members of the community at older ages.
Early on in development there is relatively little spread in the distribution because there has been relatively little opportunity as yet to either accumulate a repertory from earlier ages or to begin to effectively model behaviors characteristic of later ages. In a normally age-diverse community, the individual will however have many opportunities to broaden the range of the distribution as they mature.
I propose that the overall collective distribution function for the community as a whole operates as if it were an attractor for the individual distribution function at every age. All normal development tends to move toward a closer fit with the community, to an 'internalization of the village'. Prior to maturity, or to the modal age of the community, this pulls the peak of the distribution to the right and tends to enlarge its range and particularly to increase probabilities to either side of the maximum, with a strong tendency to induce anticipatory or precocious behavior. (Note that 'behavior' here means behavior-in-interaction as well as solo behavior, and the anticipatory domain of the distribution function corresponds closely to Vygotsky's notion of a zone of proximal development, initiated in interaction with older members of the community, particularly in responsive or reactive behavior, and later added to the solo repertory and the repertory of initiating behaviors.)
Subsequent to maturity, the community distribution would exert a 'drag' on development, and tend to retard senescence, precisely to the degree that an elder continues to have regular interactions with younger members of the community. I would expect that the elder's more youthful behavioral repertory would be maintained as responsive or reactive repertory even after it had become much less frequent as solo or initiating behavior.
Note: I refer to developmental heterochrony as 'internal heterochrony' to distinguish it from 'external heterochrony' which refers to the heterarchical connections between levels of a complex dynamical system across widely different (non-adjacent) timescales of its organizational processes. External heterochrony assumes mediation across widely different (many orders of magnitude) timescales by material-semiotic artifacts. Internal heterochrony occurs even without such phenomena, as a consequence of normal relationships between adjacent timescales. In the particular case of human development, the N+1 timescale is that of the community and change in the distribution function for the community as a whole. The N-1 and lower timescales are those for neurological processes mediating behavioral change in the organism as a whole. The focal (N) level timescale is that for observable changes in normal or typical situation-specific behaviors of the individual.
For additional discussion of the general model of ecosocial
dynamics, see papers listed on this page.