The inconvenient thing about the past is that it is impossible to set up an experiment, for instance to evoke exactly the type of utterances that are needed for a language study. Instead one has to make do with whatever source material is available. Although (self-conscious) Ulster Scots was not used for official written purposes, a good number of Ulster Scots poems from the 18th Century survive and they can be compared to Lallans and Doric poetry from around the same time as well as to modern Ulster Scots poetry and writing. But some things need to be taken into account when using poetry as a linguistic resource. The language used in poetry does not per se reflect "naturalistic" speech of the time that it was written; it is just another register with words used because they are deemed to sound poetic, or abbreviated to make the sentence run better. However, even if it is written and not recorded, poetry is a surprisingly good source for figuring out how speech would have sounded around the time of writing.
The study of Old English has learnt a lot from studying the poetry of Chaucer (1343-1400); about the vocabulary and grammar of the language but also about the way it was pronounced. The latter can be found out by comparing words that would be expected to rhyme - if Chaucer rhymes melody and eye this shows that his pronunciation of either one or the other must have differed from that today. For this reason poetry is even "likely the best source" for how historical speech sounded, says Linde Lunney (2006: 51). She notes that it is especially important to analyse Ulster Scots poetry or "the poetry written in what purports to be a written version of the Ulster-Scots spoken dialect", because "in this sub-genre the poet's awareness of the sounds of speech, and his efforts to recreate them in written form, are as important to the 'message' as they are to the 'medium'." Ulster-scots poetry usually contains nonstandard spellings that are intended to reflect how the word would actually be pronounced by Ulster-Scots speakers.
It should be noted that identifying the vocabulary from the poem is not as easy as it may seem at first glance. Words used in poetry do not always (or: often don't) reflect the "normal" vocabulary at that time as authors may choose words that sound more learned or beautiful instead of words people would usually have used. There are some periods in which it was common for poets to be very innovative with their language use. An obvious example is Shakespeare, who is credited with significantly expanding the language by using existing verbs as nouns and adjectives and vice versa, and coining a great deal of words too, for instance compulsive. (see OED: "compulsive"). Andrew Zurcher notes about studying an innovative poet such as Edmund Spenser (1552-1599) that
"it becomes very difficult to register exactly which parts of his verse qualify as innovative. What looks convincingly like archaism to us today might not, to an Elizabethan, have seemed strange at all. By contrast, words that we now accept as standard, or quasi-standard, might have seemed unfamiliarly archaic to Elizabethan readers." (Zurcher 2005)This problem is very much true for Ulster Scots poetry as well. Out of such motivations as protectivism of the language, the poets would often have sought to use as "pure" Ulster Scots as possible with the most un-english words they could find, no matter whether these would have sounded unusual to the casual speaker. A recent example, William Forbes Marshall (1888-1959) is described as "an aficionado of Ulster Scots and an avid collector of words and phrases." (**: 361)
