Grammatical and syntactic sentences

this regard must be coupled with other socio-linguistic observations. For more information on the non- comparability of different recorded texts, see Casad 1974:90. Additionally each individual speaker’s personal exposure to each variety will greatly impact hisher test scores. A distinction is often made between inherent intelligibility and acquired intelligibility. This particular survey did not screen out such participants, since its primary aim R1 was to test for extensibility of Maiyu literature to other speech communities. For this reason, those members of the community which had previously had exposure to the Maiyu variety were equally as valuable as those who had not and were therefore not barred from participating in the survey. Their inclusion in this survey is thought to provide a more accurate assessment of the community’s average comprehension of a given Berta variety, rather than only those individuals who have not travelled outside of their own variety’s area. Notably, most older Fadashi and Undulu speakers have had extensive exposure to Maiyu, but not necessarily vice versa. Furthermore, none of the Beleje Gonfoye speakers had exposure to any of the other varieties tested, and so their results present a pure reflection of inherent intelligibility levels.

3.2.3 Grammatical and syntactic sentences

3.2.3.1 Selection and creation To test for variation between the phonetic, phonological, grammatical, and syntactic structures of each speech variety surveyed, a sample of 40 sentences was constructed. These sentences were first framed in English and then translated into Amharic and were specifically designed to examine the distribution of known Maiyu structures in the larger speech area. The framework for this test was loosely based on Casad’s 1974:98 own survey among the Yawa. The sentences were intended to test for a wide range of grammatical features, juxtaposing things like transitivity, tenseaspect marking, prepositions, relative clauses, verbal compounding, demonstratives, possessives, conditionals, passives, imperatives, interrogatives, and negatives. Due to low levels of Amharic in the Undulu and Fadashi varieties, the sentence test remains incomplete. 3.2.3.2 Data sources The same people who provided the wordlist items likewise provided the oral translation of these sentences. In all cases, there was a committee of three people, one of whom was older 50+ and two of whom were of a median age 25–40. These participants had previously given consent before the elicitation of the wordlist, so no further consent was needed for this sentence survey. 3.2.3.3 Procedures For each location, immediately following the completion of the wordlist, participants were allowed a 30 minute break for tea or coffee. Once we reconvened, the committee was told an Amharic sentence and asked to translate it. Often a spokesman would give an initial answer and the other two participants would discuss and refine it. Once the committee was satisfied with their answer, they would repeat it three times for recording. Recordings were done by Gunnhild Bremer and transcribed by Nate Bremer. This procedure was followed for each sentence; however, in the case of both Undulu and Fadashi, the level of Amharic was lacking. Rather than compromise a controlled component of the test by providing translation of the sentence in another language Maiyu-Berta or Arabic, we chose to merely leave the sentence blank and the test therefore incomplete. As a result, the sentences that are recorded are believed to closely reflect the same grammatical features of the source sentence Amharic. Once we left the survey location, Nate and A. T. A. reexamined the recorded sentences and the transcriptions, and looked for variations andor mistakes in the elicitations. These observations are preserved as footnotes in Appendix K. 3.2.3.4 Analysis Analysis of the sentence test is more qualitative than quantitative. Reservations about the reliability of the results are warranted, since the test is based upon translated materials and is therefore not necessarily reflective of “natural” speech. Although undertaken in this survey, comparisons can be made with each of the RTT stories to either confirm or disprove the syntactic and grammatical constructions found in the sentence elicitation. Additionally, the phonetic and phonological structures of the sentence test’s contents can be compared with the wordlist. This test should therefore not be viewed in isolation, but rather in conjunction with other samples.

3.3 Research ethics