Minimum information required for written word recognition

Deok-Hee Kim-Dufor (1, 2) , Philippe Tigréat (1, 2) , Claude Berrou (1, 2)

1 Lab-STICC_TB_CACS_IAS
2 ELEC - Département Electronique

Reading is an automated process. One of the remarkable human abilities is that we can read even partially erased or hidden words. We carried out a study on written word recognition in order to decipher how much information is required at least to identify a word. Experimental software was designed in C++ language to measure the amount of information in pixels and reaction time. The results showed we could identify words at a very low display rate and suggest that prior knowledge on the category of words play a mediating role in written word recognition.

Mots clés: visual word recognition

Domaines: Linguistique

Deok-Hee Kim-Dufor, Philippe Tigréat, Claude Berrou. Minimum information required for written word recognition. Groupe de Linguistique Appliquée des Télécommunications (GLAT), Jun 2014, BREST, France. pp.84-91. ⟨hal-01222081⟩