Complexity in Derivational Morphology: Theory and Experimental Evidence
Lead Research Organisation:
University of Oxford
Department Name: Linguistics Philology and Phonetics
Abstract
The proposed research constitutes a combined theoretical and experimental endeavour to shed light on the question of complexity in derivational morphology and on how complexity is reflected in word processing on the evidence of brain signatures. Experimental approaches to derivation have often been hampered by overly simplistic assumptions about the structural domain of word formation (or more precisely lexeme formation, as distinct from wordform formation as in inflectional morphology). The central debate has been about whether words (lexemes) are or are not decomposed in the mental lexicon, and it has remained somewhat unclear how morphological relatedness is founded on form overlap and/or semantic compositionality, and what the role of frequencies of occurrence is in processing derived words. The planned research compares English and German, which are related languages with comparable but not identical linguistic structures. For example, suffixation and prefixation prevail and allow morpho-phonological alternations in both languages. Nevertheless, there are systematic phonological differences (e.g., stress alternations in English, umlaut alternations in German), and the morphological categories involved in derivation are not identical, either (e.g., gender is only relevant in German). There has been both behavioural as well as brain-imaging data from our own labs that suggests that the depth of morphological derivation does affect processing, but as yet we have not investigated the time-course of this effect, nor have we systematically manipulated phonological alternations in derivation. This will be investigated in a series of electrophysiological studies in German and English. We will be able to compare different types of morphological complexities, with and without phonological alternations, and our results will also allow us to make cross-linguistic generalisations in a way which is not possible when examining a single language, as is still the rule in experimental morphology.
Planned Impact
N/A
| Description | Through a combination of several experiments and links with previous studies we have found evidence in multiple languages (English and German) that when people hear/see complex words (e.g., "unhelpful", "unfreundlich"), they decompose them into their constituent parts (i.e., accessing the root "help" after processing the suffix "-ful" and prefix "un-"), but that this process is not uniform, but rather depends on the phonological (i.e., speech sound characteristics) nature of the composition. For example, when complex words are composed transparently of roots and affixes (prefixes/suffixes), with no other change to the pronunciation of the root, there is no processing cost for increases in complexity (i.e., people access "help" just as easily from "unhelpful" as they access "well" from "unwell"). This is show in both reaction times on behavioural tasks (recognising words in isolation or after being "primed" by related words) and in patterns in brain waves recorded via EEG. However, when these compositions are less transparent, such as when the root changes its phonology (sound structure; e.g., when stress shifts in CONvert vs. conVERTing, or when vowels are umlauted in German verbs) perceivers are sensitive to such effects. Again, this is reflected in both behavioural and brain data. These results are consistent with our earlier studies which showed that in the most extreme case, when changes in complexity are introduced without any clear indicators like prefixes, suffixes, or stress/vowel shifts, such as the difference between "a dress" and "to dress", listeners show the greatest processing difficulty. Therefore, we now have evidence for the first time that such characteristics which we had been able to describe formally in the past are also reflected psychologically and neurally, and this data can then be used to apply to a range of problems, such as in language learning (first and second), language attrition through aging, and in historical language change. |
| Exploitation Route | We are currently using the outcomes of our experiments (4 EEG, 8 behavioural, and 2 eye tracking) to design follow-on experiments to test the limits of listener/reader decomposition of complex words and their sensitivity to phonological (sound structure) markers of that complexity. We are also applying this work to other languages beyond English and German (e.g., Russian, Dutch, and French), and other researchers could do the same both with languages with very different word formation patterns and for different participant populations, such as second language learners, children, and clinical populations. |
| Sectors | Education Culture Heritage Museums and Collections Other |
| Description | Collaboration with the University of Konstanz on joint AHRC-DFG grant |
| Organisation | University of Konstanz |
| Department | Department of Linguistics |
| Country | Germany |
| Sector | Academic/University |
| PI Contribution | We are presently collaborating with the University of Konstanz as full partners on this joint AHRC-DFG grant: "Complexity in Derivational Morphology: Theory and Experimental Evidence". As part of this collaboration we coordinate parallel behavioural and EEG experiments in English and German, and integrate the results into shared publications and other research outputs. |
| Collaborator Contribution | Our partners at the University of Konstanz contribute on all aspects of the project, from providing input on experimental design to running experiments on German speakers in Konstanz. |
| Impact | Through this partnership we have collected parallel behavioural and EEG data on English and German word processing. This data is currently being analysed and written up for presentation/publication. |
| Start Year | 2020 |