The English Vocabulary Profile Phonological Network (EVP-Phon)
The picture at the top of my homepage is a visual representation of the L2 English Phonological network for someone at an advanced proficiency level (C2 on the Common European Framework of References for Languages, CEFR).
- Each circle represents a word.
- Each color represents the CEFR level at which students began using the word in their writing (i.e., productive usage--not receptive).
- If two words are minimal pairs (e.g., lay-lie, lay-a, lay-play), a line is drawn between those words to connect them. What you see in the picture is what happens when you draw a line between each minimal pair (i.e., phonological neighbor) in the lexicon.
EVP-Phon: A tool to analyze the L2 English mental lexicon through its phonological network
Network science tools have been used to provide critical insights into the phonological network’s structure in the mental lexicon (e.g., [1]). For example, research has shown that network structure influences word recognition [2], word learning [3], and production [4]. This work provides compelling evidence that the mental lexicon is structured and that the structure affects processing.
While great progress has been made in understanding the structure of the first language (L1) English phonological network, we still know little about the structure of the second language (L2) network. Of course, one difficulty in creating such a network is deciding which words to include. Using databases like CLEARPOND (27,751 words) [5] or KU-similarity (19,340 words) [1] are good options, but they were made to simulate a native speaker’s lexicon. The current study overcame this challenge by creating the EVP-Phon database—made specifically to simulate the L2 English phonological network. We created EVP-Phon using the English Vocabulary Profile (EVP) [6]. EVP’s goal was to identify which words learners can use –not which words they should know [7]. They did this by examining the Cambridge Learner Corpus (over 50 million words from written exams from learners all over the world) to see which words learners used at each Common European Framework of Reference (CEFR) proficiency level.
The EVP-Phon is a dynamic network that grows from the CEFR-A1 proficiency level (558 words—1034 minimal pairs) to the C2 level (6324 words—7,316 minimal pairs). Minimal pairs are defined as words that differ by adding, subtracting, or replacing one phoneme [1] (e.g., minimal pairs of “trap”: rap, strap, track, trip, etc.). The current study details the network’s structure as a whole (following [1]) as well as how that structure develops over time (i.e., by proficiency level). More specifically, we will discuss the network’s average path length, clustering coefficient, and degree distribution. We will also outline some potential ways the tool can be used to research concepts such as functional load and vocabulary acquisition. We will also show how teachers can use this tool to quickly find minimal pairs to use in their lessons and/or materials creation. Additionally, this network does not take L1 into account, but we will discuss how future iterations can do this to model specific L1-L2 pairings.
References
Network science tools have been used to provide critical insights into the phonological network’s structure in the mental lexicon (e.g., [1]). For example, research has shown that network structure influences word recognition [2], word learning [3], and production [4]. This work provides compelling evidence that the mental lexicon is structured and that the structure affects processing.
While great progress has been made in understanding the structure of the first language (L1) English phonological network, we still know little about the structure of the second language (L2) network. Of course, one difficulty in creating such a network is deciding which words to include. Using databases like CLEARPOND (27,751 words) [5] or KU-similarity (19,340 words) [1] are good options, but they were made to simulate a native speaker’s lexicon. The current study overcame this challenge by creating the EVP-Phon database—made specifically to simulate the L2 English phonological network. We created EVP-Phon using the English Vocabulary Profile (EVP) [6]. EVP’s goal was to identify which words learners can use –not which words they should know [7]. They did this by examining the Cambridge Learner Corpus (over 50 million words from written exams from learners all over the world) to see which words learners used at each Common European Framework of Reference (CEFR) proficiency level.
The EVP-Phon is a dynamic network that grows from the CEFR-A1 proficiency level (558 words—1034 minimal pairs) to the C2 level (6324 words—7,316 minimal pairs). Minimal pairs are defined as words that differ by adding, subtracting, or replacing one phoneme [1] (e.g., minimal pairs of “trap”: rap, strap, track, trip, etc.). The current study details the network’s structure as a whole (following [1]) as well as how that structure develops over time (i.e., by proficiency level). More specifically, we will discuss the network’s average path length, clustering coefficient, and degree distribution. We will also outline some potential ways the tool can be used to research concepts such as functional load and vocabulary acquisition. We will also show how teachers can use this tool to quickly find minimal pairs to use in their lessons and/or materials creation. Additionally, this network does not take L1 into account, but we will discuss how future iterations can do this to model specific L1-L2 pairings.
References
- [1] Vitevitch, M. S. (2008). What can graph theory tell us about word learning and lexical retrieval? Journal of Speech Language Hearing Research, 51, 408–422.
- [2] Chan, K. Y., & Vitevitch, M. S. (2009). The influence of the phonological neighborhood clustering coefficient on spoken word recognition. Journal of Experimental Psychology: Human Perception and Performance, 35(6), 1934.
- [3] Stamer, M. K., & Vitevitch, M. S. (2012). Phonological similarity influences word learning in adults learning Spanish as a foreign language. Bilingualism: Language and Cognition, 15(3), 490-50.
- [4] Chan, K. Y., & Vitevitch, M. S. (2010). Network structure influences speech production. Cognitive science, 34(4), 685-697.
- [5] Marian, V., Bartolotti, J., Chabal, S., & Shook, A. (2012). CLEARPOND: Cross-linguistic easy-access resource for phonological and orthographic neighborhood densities. PloS one, 7(8), e43230.
- [6] Capel, A. (2015). The English vocabulary profile. English profile in practice, 5, 9-27.
- [7] Harrison, J. (2015). What is English Profile? In J. Harrison & F. Barker (Eds.), English Profile Studies 5: English Profile in practice (pp. 1-8). Cambridge University Press.