Examine SES effects on the use of variation sets in the speech directed to children in their household’s naturalistic environment. ● Child directed speech (CDS) compared to speech between adults, shows a higher amount of repetitiveness (Newport, Gleitman & Gleitman, 1977; Söderström, 2007), in particular of naturally occurring clusters of speech with sequences of self repetitions. ● This structural phenomena, known as variation sets (Küntay & Slobin, 1996, 2002) imply partial variations in form but a constant intention and have been found to be beneficial for learning (Waterfall, 2006; Onnis et al., 2008). Why to look at the naturalistic environments? ● Previous studies focused on the behaviour of variation sets in languages such as Swedish, Croatian, English, Russian, and Hebrew (Wiren et.al, 2016; Tal & Arnon, 2018) were based on data from child-parent dyadic interactions during play in relatively brief elicited situations in the home or in the lab. ● Delimited play situations in time, space and objects are likely to elicit dense language input whereas language in everyday routines flows naturally, shows fluctuations, interspersed with silence (Bergelson, 2019), and frequent overlapping conversations among participants (Schegloff, 2000). This might affect the structural properties of CDS. Why to look at SES differences? ● Although previous findings indicated socio-economic status’ (SES) effects on the quantity of variation sets in Child Directed Speech (CDS), they were based on data from child-parent dyadic interactions in play situations (Tal & Arnon, 2018). ● SES comprises interrelated factors affecting children’s quotidianity (e.g. different levels of family education, neighborhood, household density, number of participants, etc.). These factors might relate in different ways in diverse cultural contexts (Psaki et al., 2014), and might affect CDS features differently (Pace, 2017). What effect does SES have on the quantity and extension of variation sets in the naturalistic household environment of Argentinian toddlers? Sample (Corpus: Rosemberg, Alam, Stein, Migdalek, Menti & Ojea, 2015-2016). Rosemberg, C., Alam, F., Garber, L., Stein, A., Migdalek, M. to the ● An effect of SES on the proportion of variation sets, as observed by Tal & Arnon (2018). ● The pseudo R2 is higher in the long variation sets models than in the short variation sets models. ● An age effect on the longer variation sets. ● No effect of household size on the proportion of variation sets. ● Future work will address the pragmatic function of the variation sets. Mid-SES N= 16 Age= 13.4(4.29) Maternal ed.=17.13 (3.11) Household size= 2.66(0.89) Low-SES N= 16 Age= 15.2(3.32) Maternal ed.=9.66(2.76) Household size= 5.2(2.59) Procedure ● A 4 hr audio recording was gathered with a small digital recorder worn by the child in a vest. ● The 2 hr middle were transcribed following CHAT format (64 hours). A utomatic extraction of variation sets (VS) . We developed an algorithm in Python that: -Considers a VS as two or more consecutive utter ances directed to the child that share at least one word (7, 5). -Uses MOR tier from CLAN (8) in the CHAT files to compare noun, verb and adjective lexemes in successive utterances. -Considers exact repetitions (9). -Allows: ➔ intervening utterances from the child ➔ one utterance from the speaker that did not have a word in common with the sequence of VS ➔ up to three overheard utterance from another speaker not directed to the child ➔ one intervening utterance from another speaker directed to the child. All these decisions were taken due to the specificity of naturalistic data, where several speakers participate and overlap their speech. ANALYSIS. We employed beta regression using R Statistical Models Word proportion Utterance proportion Coefficients of mean model with logit link Intercept -1.84** -2.32*** (0.69) (0.70) SES 0.52** 0.49* (0.20) (0.19) Household size 0.06 0.11 (0.08) (0.07) Age 0.08 0.11** (0.04) (0.04) Coefficients of precision model with log link Intercept 2.65* 1.13 (1.04) (1.05) SES -0.58* -0.17 (0.28) (0.29) Household size -0.05 0.17 (0.13) (0.13) Age -0.08 -0.00 (0.06) (0.06) Pseudo R2 0.19 0.26 Log Likelihood 8.12 10.48 Num. obs. 32 32 ***p < 0.001, **p < 0.01, *p < 0.05 Statistical Models Long VS, Word proportion Long VS, Utterance proportion Short VS, Word proportion Short VS, Utterance proportion Coefficients of mean model with logit link Intercept -3.27*** -3.28*** -1.27 -1.11 (0.74) (0.73) (0.78) (0.58) SES 0.46* 0.51** 0.43* 0.43* (0.19) (0.19) (0.19) (0.17) OR=1.58 OR=1.66 OR=1.54 OR=1.53 Household size -0.00 -0.04 0.16 0.06 (0.08) (0.07) (0.08) (0.07) Age 0.15*** 0.15*** 0.05 0.03 (0.04) (0.04) (0.04) (0.04) OR=1.16 OR=1.16 Coefficients of precision model with log link Intercept 2.66* 2.74* -0.72 3.59*** (1.11) (1.11) (0.99) (1.05) SES -0.55 -0.63* 0.02 -0.68* (0.30) (0.30) (0.28) (0.29) OR=0.53 OR=0.50 Household size 0.00 0.07 0.08 -0.05 (0.13) (0.14) (0.12) (0.13) Age -0.06 -0.07 0.13* -0.13* (0.06) (0.06) (0.06) (0.06) OR=1.13 OR=0.87 Pseudo R2 0.30 0.31 0.11 0.10 Log Likelihood 19.51 23.26 7.43 8.13 Num. obs. 32 32 32 32 ***p < 0.001, **p < 0.01, *p < 0.05 Examples 1 .MOT: pará que te limpio las manitos . wait for me to wash your hands . MOT: a ver la otra manito . show me your other hand . 2. MOT: se enganchó, se enganchó la tirita ahí está . it’s stuck, the strap is stuck, I got it . CHI: 0 . MOT: la tirita se había enganchado, ya está no hay problema . The strap had stuck, now there is no problem . 3. MOT: bajate de ahí hijo . get down from there, son . MOT: bajate de ahí . get down from there . MOT: bajate hijo de ahí que te vas a caer . son, get down from there, you are going to fall . MOT: Valentino bajate de ahí ! Valentino get down from there ! Socio-economic related differences in the use of variation sets in naturalistic child directed speech. A study with Argentinian population Literature cited 1. Bergelson, E., Amatuni, A., Dailey, S., Koorathota, S., & Tor, S. (2019). Day by day, hour by hour: Naturalistic language input to infants. Developmental science, 22(1), e12715. 2. Küntay, A. & Slobin, D. (1996). Listening to a Turkish mother: some puzzles for acquisition. In D. I. Slobin, J. Gerhardt, A. Kyratzis, & J. Guo (Eds.), Social interaction, social context, and language: essays in honor of Susan Ervin-Tripp (pp. 265–86). Hillsdale, NJ: Erlbaum. 3. Küntay, A. & Slobin, D. (2002). Putting interaction back into child language: examples from Turkish. Psychology of Language and Communication, 6(1), 5–1. 4. MacWhinney, B. (2000). The CHILDES Project: Tools for analyzing talk. Third Edition. Mahwah, NJ: Lawrence Erlbaum Associates. 5. Newport, E. L., Gleitman, L., & Gleitman, H. (1977). Mother, I'd rather do it myself: Some effects and non-effects of maternal speech style. In C. Snow & C. Ferguson (Eds.), Talking to children: Language input and acquisition. Cambridge: Cambridge University Press. 6. Onnis, L., Waterfall, H. R., & Edelman, S. (2008). Learn locally, act globally: learning language from variation set cues. Cognition, 109(3), 423–30. 7. Pace, A.; Luo, R.; Hirsh-Pasek, K & Golinkoff, R. (2017). Identifying pathways between socioeconomic status and language development. Annual Review of Linguistics, 3, 285-308. 8. Psaki, S. R., Seidman, J. C., Miller, M., Gottlieb, M., Bhutta, Z. A., Ahmed, T., Ahmed, A. S., Bessong, P., John, S. M., Kang, G., Kosek, M., Lima, A., Shrestha, P., Svensen, E., & Checkley W. (2014). Measuring socioeconomic status in multicountry studies: results from the eight-country MAL-ED study. Population health metrics, 12(1), 8. 9. Schegloff, E. A. (2000). Overlapping talk and the organization of turn-taking for conversation. Language in society, 29(1), 1-63. 10. Soderstrom, M. (2007). Beyond babytalk: Re-evaluating the nature and content of speech input to preverbal infants. Developmental Review, 27(4), 501-532. 11. Rosemberg, C.R., Alam, F., Stein, A., Migdalek, M., Menti, A. y Ojea, G. (2015-2016). Language Environments of Young Argentinean Children. CONICET (DOI in progress) 12. Tal, S. & Arnon, I. (2018). SES effects on the use of variation sets in child-directed speech. Journal of Child Language, 45(6), 1-16. 13. Waterfall, H. R. (2006). A little change is a good thing: feature theory, language acquisition and variation sets (Doctoral dissertation). University of Chicago.