This shows how grammatical encoding begins at functional word exchanges (message level representation and phonological encoding) and then moves to the positional level. There is evidence that shows the separation of functional and positional level processing. This evidence comes from two sources: speech error analyses and experimental data. Speech errors occur when a speaker makes semantic substitutions without planning and these substitutions are usually in the same class form. The experimental data came from lexical and syntactic priming studies. They have been able to find data that separates function and positional level processing but phonological priming can sometimes lead to a late constituent placement which may compromise the original argument (Engelhardt et al., 2009). The article was able to show two sides of the idea and showed that, although, it may be possible for grammatical encoding to be used in control structures that have no lexical content, most syntactic structures are lexically anchored. To help explain grammar encoding, Tree-Adjoining Grammar (TAG), shows that the objects of the grammar are small trees, which include a word or lexical head. When discussing the second issue of grammatical encoding being automatic or resource-demanding, the article cites a finding from Ford (1982) – which measured spontaneous speech – presented that grammatical encoding may be resource-demanding. It states that, although, grammatical encoding could be automatic, the semantic processing entails planning so it may be considered resource demanding. Limited research has been conducted to see if certain structures are more difficult to produce than others. One study cited in the article was performed by Tannenbaum and Williams (1968) in which the speakers read a story, saw a picture, and their task was to either say an active or passive
This shows how grammatical encoding begins at functional word exchanges (message level representation and phonological encoding) and then moves to the positional level. There is evidence that shows the separation of functional and positional level processing. This evidence comes from two sources: speech error analyses and experimental data. Speech errors occur when a speaker makes semantic substitutions without planning and these substitutions are usually in the same class form. The experimental data came from lexical and syntactic priming studies. They have been able to find data that separates function and positional level processing but phonological priming can sometimes lead to a late constituent placement which may compromise the original argument (Engelhardt et al., 2009). The article was able to show two sides of the idea and showed that, although, it may be possible for grammatical encoding to be used in control structures that have no lexical content, most syntactic structures are lexically anchored. To help explain grammar encoding, Tree-Adjoining Grammar (TAG), shows that the objects of the grammar are small trees, which include a word or lexical head. When discussing the second issue of grammatical encoding being automatic or resource-demanding, the article cites a finding from Ford (1982) – which measured spontaneous speech – presented that grammatical encoding may be resource-demanding. It states that, although, grammatical encoding could be automatic, the semantic processing entails planning so it may be considered resource demanding. Limited research has been conducted to see if certain structures are more difficult to produce than others. One study cited in the article was performed by Tannenbaum and Williams (1968) in which the speakers read a story, saw a picture, and their task was to either say an active or passive