Singh et al.

Topic: How bilingualism affects infants’ visual attention to novelty. Based on a Fantz (1964)–style paired-stimulus paradigm with familiar and novel images.
Age groups: $5{-}6\text{ months}$ and $8{-}9\text{ months}$ .
Locations: Spain (Basque Country) and Singapore.
Key finding: bilingual infants orient to novel information more readily than monolinguals, especially early in the task; monolinguals show a later shift toward novelty.
Implication: language experience may shape fundamental attentional preferences, not just language-specific processing.

Total participants: $N=99$ infants.
Groups: monolingual vs bilingual across two sites.
Age split: roughly half $5{-}6\text{ months}$ and half $8{-}9\text{ months}$ in each language group.
Spain: bilinguals exposed to Spanish/Basque or English/Spanish; Singapore: bilinguals exposed to English/Mandarin (and a small amount of third languages in some cases).
Sample sizes (per Table 1):
- Monolingual, 5–6 months: $N=30$ (Spain $17$ , Singapore $13$ )
- Monolingual, 8–9 months: $N=23$ (Spain $11$ , Singapore $12$ )
- Bilingual, 5–6 months: $N=24$ (Spain $13$ , Singapore $11$ )
- Bilingual, 8–9 months: $N=22$ (Spain $9$ , Singapore $13$ )
Language exposure: bilinguals have varying percentages of exposure to their two languages; third-language exposure included in Singapore sample.
Demographics: socio-economic status matched across locations; no significant differences in maternal/paternal education between sites or language groups.

Stimuli: $36$ color images of animals (cats and dogs), used to generate $24$ stimulus sequences (12 cat-only, 12 dog-only); for each sequence, $11$ images chosen per infant with $1$ familiar referent and $10$ novel referents.
Trial structure: each trial presents a pair (familiar vs novel) for $24\,s$ ; mid-trial, locations switch (to control for side bias); an attention cue appears every $12\,s$ .
Trials: each infant completed $10$ trials, across the sequences.
Recording: eye-tracking via video at 30 fps; offline coding of left/right looking via ELAN.
Data preparation: trial 10 excluded due to engagement drop; proportion looking to novel referent calculated per trial; data collapsed into $3$ blocks of $3$ trials each.
Exclusions: 2 infants excluded due to incomplete gaze data; all others included.

Primary analysis: Linear mixed-effects model (LME) predicting proportion looking to the novel referent from age group, language background, trial block, and testing location; random intercepts by participant.
Additional analyses: one-sample t-tests vs chance level $0.50$ ; exploratory correlations between novelty preference and dominant-language exposure (percent exposure to L1/L2).
Block structure: Block 1, Block 2, Block 3 (each comprising three trials).

Block 1: bilingual infants show greater novelty preference than monolinguals (significant), monolinguals near chance.
- Evidence: $\beta = -0.286, SE = 0.126, t = -2.264, p = 0.024$ (language by block interaction).
Block 2: no significant difference between groups.
- Evidence: $\beta = -0.094, SE = 0.126, t = -0.740, p = 0.460$ .
Block 3: no significant difference between groups.
- Evidence: $\beta = 0.117, SE = 0.127, t = 0.919, p = 0.359$ .
Against chance (0.50), t-tests (Bonferroni-corrected, six tests):
- Block 1: monolingual not different from chance; bilingual different from chance (p = $0.008$ ).
- Block 2: monolingual and bilingual both above chance (p = $0.023$ for monolingual; p = $0.001$ for bilingual).
- Block 3: monolingual above chance (p = $0.001$ ); bilingual not below/above at corrected level (p = $0.071$ ).

Across blocks, language exposure patterns relate to novelty, showing a shift with more balanced exposure.
Correlations (combined sample):
- Block 1: negative relation between dominant-language exposure and novelty preference.
- Block 3: positive relation between dominant-language exposure and novelty preference.
Within bilinguals (subset):
- Block 1: negative correlation, $r = -0.289, p = 0.052$ .
- Block 3: positive correlation, $r = 0.331, p = 0.025$ .
Interpretation: infants with more balanced bilingual exposure show stronger novelty orientation early, then reduced novelty orientation later, suggesting a dynamic relation between language input and attentional preference.

Bilingualism may shift attention away from familiar stimuli toward novel information, challenging the classic familiarity-to-novelty transition model (Hunter & Ames, 1988).
Two proposed mechanisms:
- Adaptive sampling in diverse linguistic environments: exposure to two languages (and associated variability) increases the need to sample broadly to learn structure, leading to greater novelty orientation.
- Increased processing efficiency or disengagement: processing demands of dual-language input may speed up familiar-information processing, accelerating transition to novelty.
Wider consequences: bilingual experience may influence general perceptual sampling and uncertainty management, not limited to language domains.
Related concepts: explore–exploit tradeoffs; heightened adaptability to uncertainty; broader sampling may support later higher-order problem solving.

Hunter & Ames (1988): three factors for familiarity-to-novelty transition; age, familiarization time, task difficulty.
Fantz (1964): foundational paradigm demonstrating stability of looking preferences with familiar vs novel stimuli.
D’Souza & D’Souza (2021); Singh (2021); Arredondo et al. (2022): bilingualism influences attention and processing in infancy and early childhood.

Bilingual exposure should be reported and modeled even in studies not measuring language directly, as it can influence looking-time measures and inferred cognitive processes.
Environmental variability (multilingual settings) can systematically shape infant attention and perception, affecting generalizability.
Results highlight the need to consider not only group differences but also continuous exposure metrics in infancy research.

Two-site study with heterogeneous populations; may not generalize to all bilingual contexts.
Whether novelty orientation is specific to bilinguals or extends to other groups with high environmental complexity remains to be tested.
Future work should include broader participant diversity and longitudinal follow-ups to assess long-term cognitive outcomes.

Bilingual exposure is associated with a greater early orientation to visual novelty in infancy, potentially due to broad sampling of a more complex linguistic environment or faster processing of familiar information.
The findings support a view of bilingualism as shaping fundamental attentional and perceptual mechanisms, with implications for broader cognitive development and methodological approaches in infant research.