Cognateness, frequency, and vocabulary size

An interactive account of bilingual lexical acquisition

Gonzalo
Garcia-Castro
Daniela S.
Ávila-Varela
Ignacio
Castillejo
Núria
Sebastian-Galles

Twitter: gongcastro

GitHub: gongcastro/isp_2023_trajectories

Word acquisition


  • Word learning is challenging: variability, ambiguity

  • Earliest evidence of word acquisition: 6 months of age

  • Bilingual word acquisition is more complex: more than one word-form per referent (gos → DOG ← perro)

  • Do bilinguals fall behind? Mixed evidence across language pairs: English-French, Catalan-Spanish, etc.

Linguistic distance


Bilingual toddlers learning two languages that share more cognates show larger vocabulary sizes

Cognates
Form-similar translation equivalents (TEs)
Cognate Non-cognate
[cat] /ˈgat-ˈga.to/ [dog] /ˈgos-ˈpe.ro/

Cognates are acquired earlier than non-cognates. Why?

Floccia et al. (2018), Mitchell, Tsui, and Byers-Heinlein (2022), Bosch and Ramon-Casas (2014), Bilson et al. (2015)

Parallel activation: candidate mechanism?

  • Lexical access is language non-selective: both languages are co-activated, even in monolingual situations
  • Cross-language activation across translation equivalents
  • Dissociation between models of bilingual word processing and word acquisition

Spivey and Marian (1999), Costa, Caramazza, and Sebastian-Galles (2000)

Accumulator models


Word acquisition as a continuous process of lexical consolidation: accumulation of word learning instances

Hidaka (2013), Mollica and Piantadosi (2017)

Simulating bilingual word acquisition

For participant i and word j:

\begin{aligned} \text{Learning instances}_{ij} &= \text{Age}_i \cdot \text{Frequency}_j \cdot \text{Exposure}_i\\ \text{Frequency}_j &\sim \text{Poisson}(\lambda) \\\\ \textbf{For simulations:}~ \lambda &= 50 \end{aligned}

Simulation 1: no parallel activation

Catalan Spanish
Language exposure 60% 40%

Catalan Spanish
Language exposure 60% 40%

Simulation 2: parallel activation


Parallel activation hypothesis
Word-representations receive learning instances from their translations. This increment in learning instances is proportional to form-similarity (cognateness).

Cognate

Non-cognate

Catalan Spanish
Language exposure 60% 40%

Catalan Spanish
Language exposure 60% 40%

Testing hypotheses

Data collection

Questionnaire

Barcelona Vocabulary Questionnaire (BVQ)

  • On-line, inspired by CDI
  • 1,600 words: 800 Catalan and 800 Spanish (sub-lists of 500 words)
  • Short-listed 302 (noun) translation pairs

Understands Understands & Says
chair [ x ] [   ]
table [   ] [   ]
… [   ] [ x ]

Fenson et al. (1994)

Participants

138,078 item responses from 366 Catalan-Spanish bilinguals

1 time 2 times 3 times 4 times
312 42 8 4


Testing hypotheses

Modelling and statistical inference

Model


Multilevel, ordinal regression model:

  • No < Understands < Understands and Says

Bayesian (brms/Stan): probability of parameter values

P(\text{model} | \text{data}) \propto P(\text{data} | \text{model}) \times P(\text{model})

Predictors


  • Age
  • Length: number of phonemes in word-form
  • Exposure: lexical frequency \times language exposure
  • Cognateness: Levenshtein similarity between a word-form and its translation

Two-way and three-way interactions between age, exposure, and cognateness

Predictor Estimate 95% HDI p(H0)
Intercepts
Comprehension and Production 0.438 [-0.5, 0.5] 0.088
Comprehension 0.936 [2.44, 0.95] 0.000
Slopes
Age (+1 SD, 4.87, months) 0.405 [1.43, 0.45] 0.000
Exposure (+1 SD, 1.81) 0.233 [0.8, 0.27] 0.000
Cognateness (+1 SD, 0.26) 0.058 [0.06, 0.1] 0.037
Length (+1 SD, 1.56 phonemes) -0.062 [-0.35, -0.04] 0.000
Age × Exposure 0.071 [0.16, 0.1] 0.000
Age × Cognateness 0.014 [0, 0.03] 0.985
Exposure × Cognateness -0.057 [-0.28, -0.05] 0.000
Age × Exposure × Cognateness -0.018 [-0.11, -0.01] 0.975

Discussion

  • Cognateness facilitates word acquisition
  • Only low-exposure words benefit from their cognate status: less dominant language receives more facilitation
  • Parallel activation as mechanism that boosts lexical consolidation: increment in cumulative learning instances
  • Next steps: word-learning, formalisation

Thanks!

  • participants and families

Appendix

Item properties

Levenshtein similarity

Levenshtein distance: number of edits for two character strings to become identical

Orthography Phonology String
Catalan porta /ˈpɔɾ.tə/ pɔɾtə
Spanish puerta /ˈpweɾ.ta/ pweɾta

Levenshtein similarity

1-\frac{lev(A, B)}{Max(length(A), length(B))}

Catalan Spanish Levenshtein
porta (/ˈpɔɾ.tə/) puerta (/ˈpweɾ.ta/) 0.50 (3)
taula (/ˈtaw.lə/) mesa* (/ˈmesa/) 0.00 (5)
cotxe (/ˈkɔ.t͡ʃə/) coche (/ˈkot͡ʃe/) 0.40 (3)
… … …

Vocabulary

Simulation: monolinguals

For participant i and word j:

\begin{aligned} \text{Learning instances}_{ij} &= \text{Age}_i \cdot \text{Frequency}_j \\ \text{Frequency}_j &\sim \text{Poisson}(\lambda) \\\\ \textbf{For simulations:}~ \lambda &= 50 \end{aligned}


For participant i and word j:

\begin{aligned} \text{Learning instances}_{ij} &= \text{Age}_i \cdot \text{Frequency}_j \\ \text{Frequency}_j &\sim \text{Poisson}(\lambda) \\ \\ \textbf{For simulations:}\\ \lambda &= 50 \\ \text{Threshold} &= 300 \end{aligned}


For participant i and word j:

\begin{aligned} \text{Learning instances}_{ij} &= \text{Age}_i \cdot \text{Frequency}_j \\ \text{Frequency}_j &\sim \text{Poisson}(\lambda) \\ \\ \textbf{For simulations:}\\ \lambda &= 50 \\ \text{Threshold} &= 300 \\ \\ \text{Age of Acquisition}_{ij} &= \text{Age}_{i~[\text{Threshold]}} \end{aligned}


For participant i and word j:

\begin{aligned} \text{Learning instances}_{ij} &= \text{Age}_i \cdot \text{Frequency}_j \\ \text{Frequency}_j &\sim \text{Poisson}(\lambda) \\ \\ \textbf{For simulations:}\\ \lambda &= 50 \\ \text{Threshold} &= 300 \\ \\ \text{Age of Acquisition}_{ij} &= \text{Age}_{i~[\text{Threshold]}} \end{aligned}


Simulations: bilinguals

Including learning instances from parallel activation:

  • Hypothesis: word-representations receive learning instances from their translations

  • Proportional to the amount of form-similarity (cognateness)

\begin{aligned} \textbf{Monolinguals:} \\ \text{Learning instances}_{ij} &= Age_i \cdot Frequency_j \end{aligned}

\begin{aligned} \textbf{Bilinguals:} \\ \text{Learning instances}_{ij} &= \text{Age}_i \cdot \text{Frequency}_j \cdot \text{Exposure}_i+ \\ &(\text{Cognateness}_j \cdot \text{Learning instances}_{ij'}) \end{aligned}

References

Bergelson, Elika, and Daniel Swingley. 2012. “At 6–9 Months, Human Infants Know the Meanings of Many Common Nouns.” Proceedings of the National Academy of Sciences 109 (9): 3253–58. https://doi.org/10.1073/pnas.1113380109.
Bilson, Samuel, Hanako Yoshida, Crystal D Tran, Elizabeth A Woods, and Thomas T Hills. 2015. “Semantic Facilitation in Bilingual First Language Acquisition.” Cognition 140: 122–34.
Bosch, Laura, and Marta Ramon-Casas. 2014. “First Translation Equivalents in Bilingual Toddlers’ Expressive Vocabulary: Does Form Similarity Matter?” International Journal of Behavioral Development 38 (4): 317–22. https://doi.org/10.1177/0165025414532559.
Costa, Albert, Alfonso Caramazza, and Nuria Sebastian-Galles. 2000. “The Cognate Facilitation Effect: Implications for Models of Lexical Access.” Journal of Experimental Psychology: Learning, Memory, and Cognition 26: 1283–96. https://doi.org/10.1037/0278-7393.26.5.1283.
Fenson, Larry, Philip S. Dale, J. Steven Reznick, Elizabeth Bates, Donna J. Thal, Stephen J. Pethick, Michael Tomasello, Carolyn B. Mervis, and Joan Stiles. 1994. “Variability in Early Communicative Development.” Monographs of the Society for Research in Child Development 59 (5): i–185. https://doi.org/10.2307/1166093.
Floccia, Caroline, Thomas D. Sambrook, Claire Delle Luche, Rosa Kwok, Jeremy Goslin, Laurence White, Allegra Cattani, et al. 2018. “I: Introduction.” Monographs of the Society for Research in Child Development 83 (1): 7–29. https://doi.org/10.1111/mono.12348.
Hidaka, Shohei. 2013. “A Computational Model Associating Learning Process, Word Attributes, and Age of Acquisition.” PLOS ONE 8 (11): e76242. https://doi.org/10.1371/journal.pone.0076242.
Hoff, Erika, Cynthia Core, Silvia Place, Rosario Rumiche, Melissa Señor, and Marisol Parra. 2012. “Dual Language Exposure and Early Bilingual Development*.” Journal of Child Language 39 (1): 1–27. https://doi.org/10.1017/S0305000910000759.
Jusczyk, P. W., and R. N. Aslin. 1995. “Infants′ Detection of the Sound Patterns of Words in Fluent Speech.” Cognitive Psychology 29 (1): 1–23. https://doi.org/10.1006/cogp.1995.1010.
Mitchell, Lori, Rachel K. Y. Tsui, and Krista Byers-Heinlein. 2022. “Cognates Are Advantaged in Early Bilingual Expressive Vocabulary Development.” PsyArXiv. https://doi.org/10.31234/osf.io/daktp.
Mollica, Francis, and Steven T. Piantadosi. 2017. “How Data Drive Early Word Learning: A Cross-Linguistic Waiting Time Analysis.” Open Mind 1 (2): 67–77. https://doi.org/10.1162/OPMI_a_00006.
Spivey, Michael J, and Viorica Marian. 1999. “Cross Talk Between Native and Second Languages: Partial Activation of an Irrelevant Lexicon.” Psychological Science 10 (3): 281–84.

International Symposium of Psycholinguistics | Vitoria, 31st May, 2023