Written cognate treatment in a Welsh-English bilingual aphasic patient

Bilingual language representation has been a great source of debate. Cognates are interesting concepts as they share meaning and have close phonology across languages (e.g the Welsh for cat is cath) so it is possible they share part of their representation. Previous spoken cognate treatment has found evidence of generalization across languages (e.g. Kohnert, 2004). There have been limited across language written treatment studies but neither of these have investigated cognates(Roberts, 2011; Tainturier, Roberts, & Roberts, 2011). It remains unknown if written treatment of cognates will also result in a cognate facilitation effect. The purpose of this study is to investigate written picture naming treatment in bilingual patients with aphasia. We will also be investigating the role of homophonolgy in cognate generalization; will generalization (if any) be equal for cognates that have exactly the same phonology (e.g. fig/ffig is pronounced /fIg/) as words that differ in phonology (e.g. cat/cath is /kæt/,/kæθ/). Procedure Participants will be people with aphasia and impaired written naming with equal proficiency in English and Welsh prior to injury. Written naming of treatment and control items will be assessed in three baseline sessions in each language 3 weeks apart to match the treatment period. A written naming treatment will be administered 3 times a week for 3 weeks and followed up with 2 post-treatment assessments also 3 weeks apart. During baselines and post-tests participants will be asked to write the name of a picture presented along with a short definition. After written naming of the stimuli participants will be asked to name the stimuli orally. Treatment will be a copy and recall treatment (e.g..Beeson, 1999) Stimuli Stimuli will consist of 8 sets of 20 items: a) treated English cognates with the same phonology (e.g. fig), b) treated English cognates with differing phonology (e.g. cat) c) treated English non-cognates (e.g. cow), d) untreated English non-cognates (e.g. table) e) untreated Welsh translations of a (e.g. ffig) f) untreated Welsh translations of b (e.g. cath) g) untreated Welsh translations of c (e.g. bwch) h) un-treated welsh translations of d (e.g. bawd). Treatment will be conducted in English and only on sets a,b,c Results Weighted Statistics (Howard, Best, & Nickels, 2015) will be used to investigate improvement from baselines to post-tests. This will establish whether there are effects of English treatment on the treated sets (a, b and c) and the untreated English and Welsh sets (d,e,f,g). If treatment improves written production of the treated sets, then we can attribute this to strengthening lexical representation and investigate generalisation to untreated sets. We will specifically be looking at whether cognates when treated in English improve in Welsh compared to non-cognates, but also if any inhibition occurs if cognates share exactly the same phonology (sets a and e) compared to if cognates have slightly differing phonology (sets c and f). The results of the study will be used to inform bilingual language models on interconnectivity and the special status of cognates as well as inform us of cost and time effective aphasia therapy.