From the lab to the field: Exploring the interplay of literacy, socioeconomic status, and the use of multiple linguistic cues on predictive processing during spoken language comprehension

Open Access
- Author:
- Velez-Aviles, Jessica
- Graduate Program:
- Spanish
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- March 25, 2024
- Committee Members:
- Paola Migliaccio-Dussias, Chair & Dissertation Advisor
John Lipski, Major Field Member
Rena Cacoullos, Major Field Member
Karen Miller, Major Field Member
Janet van Hell, Outside Unit & Field Member
Paola Migliaccio-Dussias, Program Head/Chair - Keywords:
- prediction
predictive processing
eye tracking
spoken language processing
literacy - Abstract:
- In two visual world eye-tracking experiments, this dissertation examines the role of literacy on predictive processing during spoken language comprehension. Previous studies have shown that comprehenders can use morphosyntactic cues available in the input to generate predictions about upcoming information. Recent key findings also suggest that prediction is modulated by markers of language skill such as literacy (e.g., Favier et al., 2021; Kukona et al., 2016; Mishra et al., 2012). A consistent finding in the literature is that high literacy individuals predict target information earlier than participants with low literacy levels. However, prior work has not consistently controlled for socioeconomic status (SES), an important confounding variable related to a vast array of outcomes across the lifespan, including lower educational achievement and poorer cognitive, language, and memory development (Hackman et al., 2010; Fernald et al., 2013; Morgan, 2009). Thus, the advantage in predictive processing reported in past studies for higher literacy individuals may be due to SES-related disparities and not to literacy skill per se. This dissertation investigated the role of literacy on predictive processing during spoken language comprehension with adults with varying levels of literacy who also belong to the same SES. Participants were native speakers of a vernacular variety of Spanish spoken in San Basilio de Palenque, a small Afro-Colombian village that offers a unique opportunity to keep socioeconomic status constant while varying participants’ literacy skill. The village faces many economic difficulties that impede social mobility and sources of employment are limited, especially those that could provide better living conditions. As a result, Palenqueros (the residents of San Basilio de Palenque) are from low SES and live in what would be considered substandard living conditions. Experiment 1 tested whether speakers differentially use grammatical gender information encoded in Spanish articles to anticipate upcoming nouns, and whether this ability was modulated by literacy skill. Higher and lower literacy individuals saw two-picture visual displays in which items matched or did not match in grammatical gender. Participants’ eye movements were recorded while they listened to Spanish sentences in which target items were preceded by a feminine or masculine article that agreed in gender with both pictures in the visual scene or with one of the pictures (e.g., Encuentra laFEM/elMASC… ‘Find the’). The data were analyzed by comparing the proportion of eye fixations on target objects in each condition. The results showed that higher and lower literacy speakers looked sooner at the target item on different-gender trials than on same-gender trials, replicating results from previous studies with monolingual Spanish speakers (e.g., Lew-Williams & Fernald, 2007). Crucially, both participant groups predictively looked at the upcoming target object at the earliest possible moment, suggesting that high but also low literacy individuals were able to engage in predictive processing. Experiment 2 examined the time course of predictive processing by analyzing whether speakers differentially use semantic cues presented alongside grammatical gender (morphosyntactic) information to anticipate upcoming nouns, and whether this predictive ability is modulated by literacy skill. A subset of the participants who took part in Experiment 1saw four-picture visual displays in which items were edible or not. Participants’ eye movements were recorded while they listened to Spanish sentences in which target items were preceded by: (1) a semantically constraining or non-constraining verb (e.g., comer ‘eat’ vs. comprar ‘buy’), (2) a semantically constraining partitive phrase (e.g., un pedazo de ‘a piece of’), and (3) a feminine or masculine determiner that agreed in gender with both pictures in the visual scene or with one of the pictures (Mi hermana probó/A mi hermana se le cayó un trocito de eseMASC pescadoMASC/esaFEM manzanaFEM (‘My sister tasted/dropped a piece of that fish/apple’). Each of these segments offered cues to identify the correct target. The data were analyzed by comparing the proportion of eye fixations on target objects in each condition. The results showed that both higher and lower literacy individuals engaged in anticipatory eye movements at each segment, underscoring the robustness of predictive processing across literacy levels. Overall, the results provide converging evidence for the predictive processing abilities among speakers of varying levels of literacy. The results indicate that higher and lower literacy participants are sensitive to cues in the input that provide information toward a target object. Taken together, the findings from this dissertation suggest that the role of literacy on predictive processing is far from straightforward. Under conditions where SES is carefully controlled and multiple cues support comprehenders’ predictions, the impact of literacy becomes nuanced.