The Co-evolution of Speech and the Lexicon: The Interaction of - PowerPoint PPT Presentation

The Co-evolution of Speech and the Lexicon: The Interaction of Functional Pressures, Redundancy, and Category Variation Winter & Wedel (2016) Presented by Miriam Schulz Seminar : Exemplar Theory 3 June 2020 Lecturer : Prof. Bernd Möbius

1. Introduction 2. The computational model Outline 3. Simulations 4. Conclusion

1. Introduction 2. The computational model Outline 3. Simulations 4. Conclusion

The evolution of spoken language Change Stability in in form function Miriam Schulz Winter & Wedel (2016) 3 June 2020 4

The evolution of spoken language Change Stability in in form function cot caught [k ၁ t] Miriam Schulz Winter & Wedel (2016) 3 June 2020 5

The evolution of spoken language Change Stability in in form function cot caught [k ၁ t] If language changes constantly, how can we maintain meaning? Miriam Schulz Winter & Wedel (2016) 3 June 2020 6

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) Miriam Schulz Winter & Wedel (2016) 3 June 2020 7

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) ● Word pronunciation is influenced by stored exemplars of the word , as well as of its individual sounds Miriam Schulz Winter & Wedel (2016) 3 June 2020 8

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) ● Word pronunciation is influenced by stored exemplars of the word , as well as of its individual sounds Pronunciation of ‘cot’ Miriam Schulz Winter & Wedel (2016) 3 June 2020 9

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) ● Word pronunciation is influenced by stored exemplars of the word , as well as of its individual sounds Pronunciation of ‘cot’ Stored exemplars of the word ‘cot’ Miriam Schulz Winter & Wedel (2016) 3 June 2020 10

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) ● Word pronunciation is influenced by stored exemplars of the word , as well as of its individual sounds Pronunciation of ‘cot’ Stored exemplars of Stored exemplars of the word ‘cot’ the sounds of ‘cot’: /k/, / ၁ /, /t/ Miriam Schulz Winter & Wedel (2016) 3 June 2020 11

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) ● Word pronunciation is influenced by stored exemplars of the word , as well as of its individual sounds Perception of ‘cot’ Miriam Schulz Winter & Wedel (2016) 3 June 2020 12

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) ● Word pronunciation is influenced by stored exemplars of the word , as well as of its individual sounds Perception of ‘cot’ Word-level update of ‘cot’ Miriam Schulz Winter & Wedel (2016) 3 June 2020 13

A multi-level exemplar framework ● Speech as a repeated cycle of production and perception (Pierrehumbert 2001) ● Word pronunciation is influenced by stored exemplars of the word , as well as of its individual sounds Perception of ‘cot’ Word-level update Sound-level of ‘cot’ update of /k/, / ၁ /, /t/ Miriam Schulz Winter & Wedel (2016) 3 June 2020 14

Hidden variation ● Variation in pronunciation does not always impact categorization success Miriam Schulz Winter & Wedel (2016) 3 June 2020 15

Hidden variation ● Variation in pronunciation does not always impact categorization success It doesn’t Are you saying matter if I “roca” and pronounce /k/ “roja” sound the or /x/ ! same to you?! English native Spanish native Miriam Schulz Winter & Wedel (2016) 3 June 2020 16

Hidden variation ● Variation in pronunciation does not always impact categorization success /k/ /x/ /k/ /x/ Schematized Schematized variation of /k/ variation of /k/ in English in Spanish Miriam Schulz Winter & Wedel (2016) 3 June 2020 17

Hidden variation ● Variation in pronunciation does not always impact categorization success ● Analogy of neutral or cryptic variation in biology (Wagner 2005) Miriam Schulz Winter & Wedel (2016) 3 June 2020 18

Hidden variation ● Variation in pronunciation does not always impact categorization success ● Analogy of neutral or cryptic variation in biology (Wagner 2005) Cryptic variation in biology : “variation that is not visible to evolution” Miriam Schulz Winter & Wedel (2016) 3 June 2020 19

Hidden variation ● Variation in pronunciation does not always impact categorization success ● Analogy of neutral or cryptic variation in biology (Wagner 2005) Cryptic variation in biology : “variation that is not visible to evolution” Source: Félix & Wagner (2008) Miriam Schulz Winter & Wedel (2016) 3 June 2020 20

Hidden variation ● Variation in pronunciation does not always impact categorization success ● Analogy of neutral or cryptic variation in biology (Wagner 2005) Cryptic variation in biology : “variation that is not visible to evolution” ≈ Variation in stored exemplars Source: Félix & Wagner (2008) Miriam Schulz Winter & Wedel (2016) 3 June 2020 21

Hidden variation ● Variation in pronunciation does not always impact categorization success ● Analogy of neutral or cryptic variation in biology (Wagner 2005) Cryptic variation in biology : “variation that is not visible to evolution” ≈ Variation in stored exemplars ≈ Production noise Source: Félix & Wagner (2008) Miriam Schulz Winter & Wedel (2016) 3 June 2020 22

Hidden variation ● Variation in pronunciation does not always impact categorization success ● Analogy of neutral or cryptic variation in biology (Wagner 2005) Cryptic variation in biology : “variation that is not visible to evolution” ≈ Variation in stored exemplars ≈ Production noise ≈ Word recognition Source: Félix & Wagner (2008) Miriam Schulz Winter & Wedel (2016) 3 June 2020 23

This paper Research questions: How do the distribution of word categories and the distribution of ➔ sound categories interact? Miriam Schulz Winter & Wedel (2016) 3 June 2020 24

This paper Research questions: How do the distribution of word categories and the distribution of ➔ sound categories interact? How can the system of sound categories evolve? ➔ Miriam Schulz Winter & Wedel (2016) 3 June 2020 25

This paper Research questions: How do the distribution of word categories and the distribution of ➔ sound categories interact? How can the system of sound categories evolve? ➔ Hypothesis: The more words a specific sound contrast distinguishes, the less likely that contrast is to be lost. Miriam Schulz Winter & Wedel (2016) 3 June 2020 26

This paper Research questions: How do the distribution of word categories and the distribution of ➔ sound categories interact? How can the system of sound categories evolve? ➔ Hypothesis: The more words a specific sound contrast distinguishes, the less likely that contrast is to be lost. Method: Simulation; use computational model as conceptual tool Miriam Schulz Winter & Wedel (2016) 3 June 2020 27

1. Introduction 2. The computational model Outline 3. Simulations 4. Conclusion

The computational model: setup ● Two agents Miriam Schulz Winter & Wedel (2016) 3 June 2020 29

The computational model: setup ● Two agents ● Each with a mental lexicon ○ e.g. a 4-word-lexicon: { ba, pa, bi, pi } Miriam Schulz Winter & Wedel (2016) 3 June 2020 30

The computational model: setup ● Two agents ● Each with a mental lexicon ○ e.g. a 4-word-lexicon: { ba, pa, bi, pi } Every word is seeded with some exemplars ● ○ e.g. ba: [ ba 1 , ba 2 , …, ba n ] Miriam Schulz Winter & Wedel (2016) 3 June 2020 31

The computational model: setup ● Two agents ● Each with a mental lexicon ○ e.g. a 4-word-lexicon: { ba, pa, bi, pi } Every word is seeded with some exemplars ● ○ e.g. ba: [ ba 1 , ba 2 , …, ba n ] ● Each exemplar varies along two continuous phonetic dimensions ○ Let’s assume that dimension 1 = voicing (/b/ vs. /p/); dimension 2 = vowel height (/a/ vs. /i/) Miriam Schulz Winter & Wedel (2016) 3 June 2020 32

The computational model: setup ● Two agents ● Each with a mental lexicon ○ e.g. a 4-word-lexicon: { ba, pa, bi, pi } Every word is seeded with some exemplars ● ○ e.g. ba: [ ba 1 , ba 2 , …, ba n ] ● Each exemplar varies along two continuous phonetic dimensions ○ Let’s assume that dimension 1 = voicing (/b/ vs. /p/); dimension 2 = vowel height (/a/ vs. /i/) ○ /ba/ ≈ (30,30) vs. /pi/ ≈ (70,70) Miriam Schulz Winter & Wedel (2016) 3 June 2020 33