Simrat Binning¹, Ipek Oruc¹; ¹University of British Columbia

Blur disrupts visual recognition by reducing high spatial-frequency information needed for fine-detail discrimination. Because distant objects project blurrier retinal images, recognition is expected to worsen with distance. However, our previous work on faces revealed a different pattern: blurry faces are recognized more accurately when viewed small or afar than when large or near - an effect we termed the blur paradox. Whether this phenomenon is face-specific or a general visual mechanism remains unclear. Here, we tested whether it extends to word recognition. We selected 100 frequency-matched words from the Corpus of Contemporary American English. Intact and blurry versions were generated at four sizes (81, 270.5, 460, 650 pt), using a Butterworth lowpass filter (cut-off: 1.1 cycles/letter). Stimuli were presented at a distance of 114 cm. Participants (Expt. 1: n=12; Expt. 2: n=30) completed 100 randomly ordered trials, viewing each word once in one of four conditions. Experiment 1 used a 2 x 2 within-subjects design: size (650, 81 pt) x blur (intact, blurry). Experiment 2 presented only blurry words at one of four sizes. Percent correct recognition was recorded. Recognition was at ceiling in both intact conditions (Expt. 1). Blur significantly reduced accuracy, demonstrated by a main effect (F(1,11)=156.8, p<0.0001). A main effect of size F(1,11)=170.7, p<0.0001) was also qualified by a size x blur interaction (F(1,11)=170.7, p<0.0001), with lower accuracy for blurry words at the large size (36.7%, SD=16.3) compared to small (95.7%, SD=5.5). In Experiment 2, accuracy declined with increasing size (F(3,87)=84.63, p<0.0001), ranging from 88.9% (81pt) to 41.7% (650 pt). These results show that the blur paradox extends beyond faces, also occurring for words. Unlike face perception, small blurry words showed a near-complete recovery of recognition, indicating stimulus-specific processing strategies. These findings have implications for reading and everyday perception, where the visual system compensates for optical limitations.

Acknowledgements: This work was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant RGPIN-2025-05239 (IO) and an NSERC USRA award (SB).