Final Report Summary - MORPHOLOGY (Morphological Complexity: Typology as a Tool for Delineating Cognitive Organization)
Many languages express different grammatical meanings through inflectional morphology, as in English walk, walk-s, walk-ed and walk-ing. Inflection is one of the most uniquely human aspects of language: various systems of animal communication may be said to have a lexicon and a syntax, but only human language has inflectional morphology. Yet far from being just a tool in the communicative apparatus of language, inflectional morphology may add a degree of complexity which is at cross-purposes with the meaning or function of the utterance. This complexity, though often idiosyncratic, is highly structured; it is determined by – and hence it is a reflection of – the human mind.
The aim of the MORPHOLOGY project has therefore been to advance our understanding of morphological complexity along two dimensions: internal to the word and across sets of words. Internal to the word we have investigated the structure of the paradigm. In a straightforward system, morphological forms reflect their meaning; if a language distinguishes gender (e.g. masculine and feminine) and number (e.g. singular and plural), then a transparent paradigm would divide things four ways: masculine singular, feminine singular, masculine plural and feminine plural. A complex system comes about when a language makes odd divisions that do not appear to reflect anything about meaning, as with the Somali word for ‘this’ which has just two forms: -ta means ‘this (feminine)’ or ‘those (masculine)’, while -ka means ‘this (masculine)’ or ‘those (feminine)’. The second dimension we have investigated is across sets of words. In a straightforward system the same grammatical meaning is expressed in the same way wherever it occurs. That would be the case in English if every plural were formed by adding -s. But we also have words like women, phenomena and sheep that show that there is more to know than one simple pattern. Even so, English is relatively simple, and in some languages the separate patterns that must be learnt run into the many hundreds.
We have achieved this advance in understanding by a large-scale cross-linguistic typological survey which has amassed and analysed data on an unprecedented scale, surveying the known language families of the world, and concentrating in depth on selected hotbeds of complexity. Besides using these data as the essential bedrock for our own research, we have made the findings freely available online, by encoding them in a database, which allows users to explore the patterns manifested both within words and across sets of words. These data were naturally applied to the four investigative strands that ran through the project: (i) the theory of grammar, (ii) the history of language, (iii) psychology, and (iv) computation.
We have contributed to understanding the role of morphological complexity in grammatical theory by identifying understudied or previously unnoticed aspects of morphological structure, embedding them within a typological framework and providing them with rigorous formal analyses. These two go hand in hand, since our development of the Canonical Typology framework has clarified the issues for the formal analysis. We have traced the historical development of morphological complexity through case studies of sets of related languages whose diversity reflects alternative developmental pathways. Contrary to what is often assumed, arbitrary morphological patterns, once embedded within the system, are seldom subject to simplification, but rather persist and provide a basis for further elaboration. We have designed and conducted a series of psycholinguistic experiments that, for the first time, have investigated the influence of paradigmatic configuration in learning. In assessing and modelling morphological complexity we have made extensive use of computational methodology, implementing and verifying our theoretical proposals and exploring hypotheses about the learning of morphological systems.
For each of the specific problems tackled we have taken a step forward, unravelling the complexity of the different phenomena. We have shown more of the great range of complexity which morphological systems display, from the very simple (like English) to the dramatically complex (like Nuer and Archi). But with our greater understanding the sense of wonder is not diminished – rather it increases. The complexity we have revealed is all “backgrounded” for speakers: the most complex systems are handled effortlessly by the human mind, as speakers get on with their everyday use of their languages - communicating in different modalities for numerous purposes.
The aim of the MORPHOLOGY project has therefore been to advance our understanding of morphological complexity along two dimensions: internal to the word and across sets of words. Internal to the word we have investigated the structure of the paradigm. In a straightforward system, morphological forms reflect their meaning; if a language distinguishes gender (e.g. masculine and feminine) and number (e.g. singular and plural), then a transparent paradigm would divide things four ways: masculine singular, feminine singular, masculine plural and feminine plural. A complex system comes about when a language makes odd divisions that do not appear to reflect anything about meaning, as with the Somali word for ‘this’ which has just two forms: -ta means ‘this (feminine)’ or ‘those (masculine)’, while -ka means ‘this (masculine)’ or ‘those (feminine)’. The second dimension we have investigated is across sets of words. In a straightforward system the same grammatical meaning is expressed in the same way wherever it occurs. That would be the case in English if every plural were formed by adding -s. But we also have words like women, phenomena and sheep that show that there is more to know than one simple pattern. Even so, English is relatively simple, and in some languages the separate patterns that must be learnt run into the many hundreds.
We have achieved this advance in understanding by a large-scale cross-linguistic typological survey which has amassed and analysed data on an unprecedented scale, surveying the known language families of the world, and concentrating in depth on selected hotbeds of complexity. Besides using these data as the essential bedrock for our own research, we have made the findings freely available online, by encoding them in a database, which allows users to explore the patterns manifested both within words and across sets of words. These data were naturally applied to the four investigative strands that ran through the project: (i) the theory of grammar, (ii) the history of language, (iii) psychology, and (iv) computation.
We have contributed to understanding the role of morphological complexity in grammatical theory by identifying understudied or previously unnoticed aspects of morphological structure, embedding them within a typological framework and providing them with rigorous formal analyses. These two go hand in hand, since our development of the Canonical Typology framework has clarified the issues for the formal analysis. We have traced the historical development of morphological complexity through case studies of sets of related languages whose diversity reflects alternative developmental pathways. Contrary to what is often assumed, arbitrary morphological patterns, once embedded within the system, are seldom subject to simplification, but rather persist and provide a basis for further elaboration. We have designed and conducted a series of psycholinguistic experiments that, for the first time, have investigated the influence of paradigmatic configuration in learning. In assessing and modelling morphological complexity we have made extensive use of computational methodology, implementing and verifying our theoretical proposals and exploring hypotheses about the learning of morphological systems.
For each of the specific problems tackled we have taken a step forward, unravelling the complexity of the different phenomena. We have shown more of the great range of complexity which morphological systems display, from the very simple (like English) to the dramatically complex (like Nuer and Archi). But with our greater understanding the sense of wonder is not diminished – rather it increases. The complexity we have revealed is all “backgrounded” for speakers: the most complex systems are handled effortlessly by the human mind, as speakers get on with their everyday use of their languages - communicating in different modalities for numerous purposes.