The results of this validation study suggested that the instrument is not reliable for the analysis of reference. The qualitative data revealed that the taxonomy is not always explicit enough about the boundaries of categories and the types of antecedents particular cohesive items may have. Furthermore, it was suggested that the several uses of the definite article could be distinguished; where exophoric or grammatical types need to be specified and cohesive types need to be supplemented with the consideration of the lexical aspect. The status of full noun phrases in reference chains was also questioned, however, while carrying out lexical cohesion analysis and reference analysis with a single analytical tool might be tangible, this would lead much beyond the scope of the present study. It was also pointed out, that for a valid analysis of cohesion, a qualitative analysis is necessary. As a consequence of all this, the taxonomy could be revised in such a way as to allow for more options: to be able to show multiple presupposed items; to involve grammatical cohesion, which does take part in longer chains of ties; and to incorporate findings of recent models of text interpretation. The results of this validation study suggest that the instrument is not reliable for the analysis of reference. The qualitative data revealed that the taxonomy is not always explicit enough about the boundaries of categories and the types of antecedents particular cohesive items may have. Furthermore, it was suggested that the several uses of the definite article could be distinguished; where exophoric or grammatical types need to be specified and cohesive types need to be supplemented with consideration of the lexical aspect. The status of full noun phrases in reference chains was also questioned; however, while carrying out lexical cohesion analysis and reference analysis with a single analytical tool might be tangible, this would be much beyond the scope of the present study. It was also pointed out that for a valid analysis of cohesion, a qualitative analysis is necessary. As a consequence of all this, the taxonomy could be revised in such a way as to allow for more options: to be able to show multiple presupposed items; to involve grammatical cohesion, which does take part in longer chains of ties; and to incorporate findings from recent models of text interpretation.