“Emergence in Cognitive Science”, 2010-09-14 ():
The study of human intelligence was once dominated by symbolic approaches, but over the last 30 years an alternative approach has arisen.
Symbols and processes that operate on them are often seen today as approximate characterizations of the emergent consequences of sub-symbolic or non-symbolic processes, and a wide range of constructs in cognitive science can be understood as emergents. These include representational constructs (units, structures, rules), architectural constructs (central executive, declarative memory), and developmental processes and outcomes (stages, sensitive periods, neurocognitive modules, developmental disorders). The greatest achievements of human cognition may be largely emergent phenomena.
It remains a challenge for the future to learn more about how these greatest achievements arise and to emulate them in artificial systems.
…4.4.1. Stages and sensitive periods: Development is clearly not a completely continuous process. Piaget, of course, was famous for identifying developmental stages (1963, The developmental psychology of Jean Piaget), and although the broad stages that he envisioned have not held up, there remain good reasons to believe that children’s cognitive abilities do not advance in a completely gradual and continuous fashion. Relatedly, there appear to be sensitive periods in a wide range of domains, including vision (ocular dominance) and language (especially for syntactic and phonological aspects of language if not for other aspects).
Just what are the factors that are responsible for these effects? Recent approaches based on connectionist models have provided a way of seeing stage-like progressions as possible emergent consequences of a gradual learning process. In the early days of distributed connectionist models, I considered the developmental progress children make on a Piagetian task called the balance scale task (1976). In this work, I found that multilayer networks undergo accelerations and decelerations, exhibiting stage-like effects (1989). This work remains controversial; as a model that shares features with many other emergentist models, the transitions in it are not in fact completely abrupt; furthermore, around transitions in particular, performance in the model is graded and only ~characterizable as characteristic of the stages others have seen in children’s behavior.
In recently revisiting these issues (Schapiro & 2009), we found renewed support for the view I have held from the outset, namely that on close inspection of the data, there is evidence in children of exactly the kinds of graded effects that are seen in the model. Indeed, even stage theorists now speak in terms of “overlapping waves” instead of discrete transitions between stages (2002).
Critical periods in development (and subtler phenomena, including age of acquisition effects) are another area where emergence-based approaches have received considerable attention. A wide range of different ways of thinking about the basis of sensitive periods has been considered. Many modelers have proposed that reduced plasticity might not reflect a biological switch, but might be an emergent consequence of the accumulated effects of earlier experience (1995; 1986; Vallabha & 2007; 2004). Similarly, 2007 has shown how the vocabulary spurt in child development could reflect the simple cumulative consequences of experience. Though not quite a critical period phenomenon, it is also worth noting the work of et al 1984 on the disappearance of stepping behavior in infancy, which offers an emergentist alternative to the standard notion that this behavior disappears because of maturation of top-down inhibitory circuits. This work played a seminal role in the further development of dynamical systems approaches to development (2003; 1996).