Year: 2019
Type: dissertation
Abstract: Neighbourhood density refers to the number of words that can be derived from a \ngiven word by changing a single phoneme (phonological neighbours) or letter \n(orthographic neighbours). Performance in verbal short-term memory (vSTM) tasks \n(e.g., serial recall/reconstruction) is usually best when to-be-remembered words are \nfrom dense rather than sparse neighbourhoods. This is typically used as evidence of \nshort-term storage being supported by networks of long-term lexico-phonological \nknowledge. The primary aim of the thesis is to assess our current understanding of \nneighbourhood density and to consider what it really reveals about the nature of \nvSTM. This was achieved by adopting several approaches. Chapter 2 first identified \nand then investigated three key variables – modality, task type and word pool size – \nto assess the parameters required for a dense neighbourhood advantage in vSTM \ntasks to manifest. However, across 4 experiments, it was revealed that \nneighbourhood density does not have robust and general effects upon vSTM with an \nadvantage for sparse neighbourhood words elicited in Experiments 3 and 4. The \nfindings raise questions over several models of vSTM. Chapter 3 more closely \nexamined the parameters used by some previous experiments that have \ninvestigated neighbourhood density and found that the distribution of onset letters \nwithin each word pool is often not controlled. Simulations were used to demonstrate \nthat this oversight can produce analogous effects in serial reconstruction tasks if \nparticipants were to use part of the word (e.g., the onset letter) to inform the order of \nthe originally presented sequence. If, in some instances, the distribution of onset \nletters within word pools was determining vSTM task accuracy then the usefulness of \nusing those results to make specific claims about how neighbourhood density \nimpacts vSTM is brought into question. Chapter 4 considered that neighbourhood \ndensity variations might exist because of effort minimisation during the development \nand evolution of language. It was found that words from denser neighbourhoods tend \nto consist of more effortful articulations and take longer to vocalise than words from \nsparser neighbourhoods. This raises the possibility that neighbourhood density \ndistributions impact vSTM because dense neighbourhood words are generally easier \nto articulate, rather than dense neighbourhood words being better supported by \nnetworks of long-term lexico-phonological knowledge. Finally, Chapter 5 attempted \nto demonstrate that effects analogous to those found when neighbourhood density is \nmanipulated can also be found when only articulatory difficulty is manipulated. \nHowever, articulatory difficulty failed to predict the outcome of three experiments. \nThe results of the thesis are considered in relation to several models of vSTM and \nthe concept of neighbourhood density and what role, if any, it plays in vSTM is \ncritically discussed.
Author: Greeno, David James
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
Funder: Cardiff University
OpenAlex ID: W3036521990