Methodological considerations. There is optionality in agreement in ST, BUT agreement as a whole is not optional. o Instead of concluding that optionality of agreement holds across the board, we explored possible conditioning factors ▪ We found that the distinction between Specifier vs. Complement base position is the determining factor. o This subregularity proved to be a theoretically relevant and informative point. • The pattern proved to be persistent and robust for a group of speakers we worked with. o However, we also encountered inter-speaker variation. o Pooling the judgments across speakers would have prevented us from uncovering the pattern. Takeaway Apparent optionality may disguise hidden regularities. References Xxxxxx, X. (1987). Tzotzil clause structure. Reidel: Dordrecht. Xxxxxx, X. (1999). External Possessor and Logical Subject in Tz’utujil. In External Possession, ed. by X. Xxxxxx and X. Xxxxx. Amsterdam: Xxxxxxxxx. 451-485. Aissen, J. (2003). Differential object marking: Iconicity vs. economy. Natural Language & Linguistic Theory, 21(3), 435- 483. Xxxxxx, X. (2017). Correlates of ergativity in Mayan. In X. Xxxx, X. Xxxxxx and X. Xxxxxx eds., Handbook of Ergativity. OUP. Xxxxx, X. (1996). The Polysynthesis Parameter. Oxford University Press. Xxxxxxx, X., & Xxxx,
Methodological considerations. 5.2.1 The limitations of the approaches taken.
Methodological considerations. 5.2.1 The limitations of the approaches taken There are many limitations to this research. Analysis of qualitative studies inevitably employs a degree of subjectivity and therefore one must be cautious to assert or accept the generalizability of the findings. The potential problem of generalizability is increased as two of the studies (the focus groups’ and modified crisis information studies) included service users or carer participants or professionals that had involvement with clinical services in South London (where the author is a clinician). Two of the studies involved participants with ID and so some of the limitations in those studies are common to all research involving participants with ID, which has been historically problematic (Xxxxxxxx et al. 2011). Difficulties have included obtaining consent from participants with ID, finding services willing to participate, choosing relevant outcomes, agreeing on definitions and recruitment of participants (Xxxxxx et al. 2002). The modified crisis information feasibility study was further limited by its sample size and lack of control group. The crisis information study only included participants who understood English, which might limit how well its findings could be generalized. Those who do not speak English could be the very group in which crisis information might have the most impact. Also those recruited were more likely to be those living in staffed care homes rather than living more independently. However those who might benefit most from carrying crisis information could be those less likely to be accompanied by staff.
