Personal tools
You are here: Home Events ANC Workshop: Richard Shillcock and Matthias Hennig, Chair: Nigel Goddard

ANC Workshop: Richard Shillcock and Matthias Hennig, Chair: Nigel Goddard

— filed under:

  • ANC Workshop Talk
When Mar 05, 2019
from 11:00 AM to 12:00 PM
Where IF 4.31/33
Add event to calendar vCal

Richard Shillcock


Exploring visual-phonological systematicity in the Roman alphabet and other alphabets


[Based on Jee, Tamariz and Shillcock, submitted]


Linguists since Saussure have traditionally asserted that the relation between the form of words and their meaning is completely arbitrary. In fact, it turns out that there is a small but significant systematicity between phonological form and meaning, with considerable recent research on this issue (Blasi, Wichmann, Hammarstrom, Stadler, & Christiansen, 2016; Dautriche, Mahowald, Gibson, & Piantadosi, 2017; Monaghan, Shillcock, Christiansen, & Kirby, 2014; Shillcock, Kirby, McDonald, & Brew, 2001; Tamariz, 2008). Words that sound alike tend to have similar meanings, across a large lexicon, even after morphological structure and other historical issues have been allowed for. The effect size is very small, but it is of psychological interest.


We have extended this paradigm to the question of whether there is systematicity between the visual form of individual letters and their canonical pronunciation. There are potential implications font design, for learning to read, for ameliorating dyslexia, for interventions to simplify alphabets, and for understanding the processing involved in skilled reading. 


The systematicity of the representation of tongue position in Korean orthography (Hangeul) is well known—it was intentionally created that way. But what about the historically older Roman alphabet? Pitman’s shorthand? The ‘rational’ Shavian alphabet? Klingon?



Matthias Hennig -


Standardised access: a new approach for reproducible (neuro)science


Despite the rapid development of experimental methodology and technology, lack of experimental reproducibility is considerably slowing down progress in the biological and biomedical sciences. This is, in no small part, due to poor standardisation and documentation of often hand-drafted data analysis pipelines (in plain English: hacks). I will present a new approach we developed to assist reproducible analysis of extracellular electrophysiological recordings. Unlike previous attempts, which largely aimed at creating unifying data formats, we standardise the access to relevant aspects of the recorded data regardless of the data source. This enables, for instance, the analysis of the same data set with different software packages with different software packages, and a comparison of the outcomes, with a just few lines of code. Looking forward, I will discuss how automated provenance capture could allow fully reproducible analysis.


Joint work with Cole Hurwitz, Alessio Buccoino (Oslo) and Jeremy Magland (Flatiron, NYC)