A really nice story from Raqaiya Hasan (2006, 30):
Like many, I first encountered grammar in the context of being taught language, both Urdu, my mother tongue, and English, a second language. Now, the Urdu grammar did not bother me, and thinking back on it I can now guess why: its terminology was functional. So if it said that a jumla (roughly a clause) was khabriya (literally 'news-giver') then I knew that in the clause there will be some sort of 'news' about something or someone: usually the 'news' consisted of classifying or identifying information; so in terms of SFL (Systemic Functional Linguistics - RMcC) jumla khabriya was a relational clause, with some entity (called ism-e-zaat, literally 'noun/name of individual or class') classified or identified by reference to something (called ism-e-sifat, literally 'noun/name of quality'). Examples offered by Misbah-ul-Qawaed are white is my favourite colour and Zaid is a human being. But the grammar of English was considerably different: the language of description was entirely formal and the terms were neither semanticized nor put in the context of the use of that language. There was an English language and there was an English grammar, and they seemed to have nothing to do with each other. So I ignored all the inconsequentialities the teachers brought into their discourse by way of teaching grammar such as gerund, infinitive, adjectival complement, past participle and so on. It's not surprising that I came to linguistics completely convinced that grammar was a profitless pastime.
(Halliday, M. & Hasan, R. 2006. 'Retrospective on SFL and Literacy' Language and Literacy: Functional Approaches, edd. R. Whittaker, M. O'Donnell and A. McCabe, Continuum, London, pp. 15-44).
