(A synthesis of the two positions)
Donald O. Hebb (1904-1985), neuropsychologist (The Organization of Behavior, 1949) hypothesised that coordinated activity of a presynaptic terminal and postsynaptic neuron would strengthen the connection between them. Hebb's postulate was originally formulated to explain learning and memory but has now (post Hubel and Weisel) been applied to situations involving long-term modification in synaptic strength including those occuring during cortical development. It implies that synaptic terminals strengthened by correlated activity will be retained or sprout new branches while those persistently weakened by uncorrelated activity will eventually lose their hold on the postsynaptic cell.
Hebb argued that behavioural patterns such as visual perceptions are built up gradually over long periods of time through the connection of particular sets of cells called cell assemblies. Over time, more complex behaviours are formed out of sets of cell assemblies which he called phase sequences. Phase sequences are less localised, they involve some equipotentiality - i.e. alternative pathways which are called into use in case some are destroyed. In one sense therefore development proceeds from localisation to holism. In another sense, since plasticity is more in earlier life, development also leads to more localisation.
Since then particular kinds of learning have been shown to involve the development of particular circuits of neurons. Localisation is common at the level of about 100s of neurons.