<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0102">
  <Title>Non-locality all the way through: Emergent Global Constraints in the Italian Morphological Lexicon</Title>
  <Section position="8" start_page="2" end_page="2" type="concl">
    <SectionTitle>
6 Conclusion and future work
</SectionTitle>
    <Paragraph position="0"> The paper offered a series of snapshots of the dynamic behaviour of a Kohonen map of the mental lexicon taken in different phases of acquisition of the Italian verb system. The snapshots consistently portray the emergence of global ordering constraints on memory traces of inflected verb forms, at different levels of linguistic granularity.</Paragraph>
    <Paragraph position="1"> Our simulations highlight not only morphologically natural classes of input patterns (reminiscent of the hierarchical clustering of perceptron input units on the basis of their hidden layer activation values) and selective specialisation of neurons and prototype vector dimensions in the map, but also other non-trivial aspects of memory organisation.</Paragraph>
    <Paragraph position="2"> We observe that the number of neighbouring units involved in the memorisation of a specific morphological class is proportional to both type frequency of the class and token frequency of its members. Token frequency also affects the entrenchment of memory areas devoted to storing individual forms, so that highly frequent forms are memorised in full, rather than forming part of a morphological cluster.</Paragraph>
    <Paragraph position="3"> In our view, the solid neuro-physiological basis of SOMs' processing strategies and the considerable psycho-linguistic and linguistic evidence in favour of global constraints in morphology learning make the suggested approach an interesting medium-scale experimental framework, mediating between small-scale neurological structures and large-scale linguistic evidence. In the end, it would not be surprising if more in-depth computational analyses of this sort will give strong indications that associative models of the morphological lexicon are compatible with a &amp;quot;realistic&amp;quot; interpretation of morpheme-based decomposition and access of inflected forms in the mental lexicon. According to this view, morphemes appear to play a truly active role in lexical indexing, as they acquire an increasingly dominant position as local attractors through learning. This may sound trivial to the psycholinguistic community. Nonetheless, only very few computer simulations of morphology learning have so far laid emphasis on the importance of incrementally acquiring structure from morphological data (as opposed - say - to simply memorising more and more input examples) and on the role of acquired structure in lexical organisation. Most notably for our present concerns, the global ordering constraints imposed by morphological structure in a SOM are the by-product of purely local strategies of memory access, processing and updating, which are entirely compatible with associative models of morphological learning. After all, the learning child is not a linguist and it has no privileged perspective on all relevant data. It would nonetheless be somewhat reassuring to observe that its generalisations and ordering constraints come very close to a linguist's ontology. null The present work also shows some possible limitations of classical SOM architectures. The propensity of SOMs to fully memorise input data only at late learning stages (in the fine-tuning phase) is not fully justified in our context. Likewise, the hypothesis of a two-staged learning process, marked by a sharp discontinuity at the level of kernel radius length, has little psycholinguistic support. Furthermore, multiple classifications are only minimally supported by SOMs. As we saw, a paradigm-based organisation actually replaces the original lexical structure. This is not entirely desirable when we deal with complex language tasks. In order to tackle these potential problems, the following changes are currently being implemented: * endogenous modification of radius length as a function of the local distance between the best matching prototype vector and the current stimulus; the smaller the distance the smaller the effect of adaptive updating on neighbouring vectors * adaptive vector-distance function; as a neuron becomes more sensitive to an input pattern, it also develops a sensitivity to specific input dimensions; differential sensitivity, however, is presently not taken into account when measuring the distance between two vectors; we suggest weighting vector dimensions, so that distances on some dimensions are valued higher than distances on other dimensions * &amp;quot;self-feeding&amp;quot; SOMs for multiple classification tasks; when an incoming stimulus has been matched by the winner unit only partially, the non matching part of the same stimulus is fed back to the map; this is intended to allow &amp;quot;recognition&amp;quot; of more than one morpheme within the same input form * more natural input representations, addressing the issue of time and space-invariant features in character sequences.</Paragraph>
  </Section>
class="xml-element"></Paper>