draft manuscript - Linguistics - University of California, Berkeley
draft manuscript - Linguistics - University of California, Berkeley
draft manuscript - Linguistics - University of California, Berkeley
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
central vowel /1/. The high central vowel is generally represented as , although a less common<br />
variant, , also surfaces, particularly following the bilabial stop /p/. With this representational<br />
choice this phoneme is thus always conflated with other vowels, and we must rely on our knowledge<br />
<strong>of</strong> the corresponding form in modern Omagua to infer the appropriate phonemic representation in<br />
forms that exhibit orthographic in the Omagua texts. Additionally, the rounded back vowel<br />
/u/ is variably represented as either or , presumably due to the fact that the single<br />
rounded back vowel in the language occupied a position between cardinal /u/ and /o/, leading to<br />
variation in how the Jesuit authors perceived and represented the segment.<br />
A quite different challenge for accurate phonemic representation arises from what we assume to<br />
be errors introduced in the process <strong>of</strong> copying the text. Given the nature <strong>of</strong> the errors, we believe<br />
that the texts were copied at least once by someone who had no knowledge <strong>of</strong> Omagua. The result<br />
was scribal errors that are easily explained if we assume that the copyist was simply attempting<br />
to interpret and reproduce hand-written characters based on their shape, without being able to<br />
rely on wider knowledge <strong>of</strong> the Omagua lexicon or Omagua phonotactics. 112 Thus, for example,<br />
it is evident that orthographic handwritten was misinterpreted as or on several<br />
occasions, where both <strong>of</strong> the latter resemble (see Table 3.3 113 ).<br />
The final issue we consider with relation to the phonemic representation <strong>of</strong> forms in the Old<br />
Omagua texts is the role <strong>of</strong> sound change. As indicated in §3.1, there are few systematic differences<br />
between the phonemic representation <strong>of</strong> Old Omagua forms and modern Omagua ones. A<br />
notable exception to this generalization involves Old Omagua orthographic . Most instances<br />
<strong>of</strong> correspond to modern Omagua /I/, which is a reflex <strong>of</strong> Proto-Omagua-Kokama *e, also<br />
corresponding to Kokama-Kokamilla /e/ (O’Hagan and Wauters 2012). In light <strong>of</strong> this, is<br />
most typically represented as /e/ in our analysis in the line <strong>of</strong> phonemic representation (see (3.1c)).<br />
More infrequently, however (i.e., only in the Lord’s Prayer and Catechism Fragment), corresponds<br />
to modern /i/, particularly in unstressed position, and crucially also to Kokama-Kokamilla<br />
/i/. In these instances, Old Omagua data becomes crucial for reconstructing the proper Proto-<br />
Omagua-Kokama segment, since we would otherwise reconstruct *i in such forms. In the majority<br />
<strong>of</strong> these instances, comparative data from elsewhere in the Tupí-Guaraní family has shed light on<br />
this particular issue, in that Old Omagua words that exhibit but correspond to both modern<br />
Omagua and Kokama-Kokamilla /i/ typically correspond to /e/ in other languages <strong>of</strong> the family.<br />
In these cases we represent again as /e/ in our phonemic representation, and note that this<br />
yields an additional correspondence set between Old Omagua and modern Omagua (really between<br />
Proto-Omagua-Kokama and both daughter languages) e:i.<br />
Conversely, and even less frequently, corresponds to modern /I/, which clearly came from<br />
Old Omagua /e/. This latter correspondence, which is the least well attested, suggests to us that<br />
the /e/ phoneme in some words may have already been raising to /I/ at this time, resulting in<br />
112 For the Lord’s Prayer and Catechism Fragment, the copying event was likely only that <strong>of</strong> typesetting the texts for<br />
publication, since, although there are numerous unexpected graphemes in the forms <strong>of</strong> these texts, word breaks are<br />
more faithfully reproduced. However, in the Full Catechism and Pr<strong>of</strong>ession <strong>of</strong> Faith, which come to us bundled in<br />
the appendices to Manuel Uriarte’s diaries (see §6.1), word breaks are additionally quite surprising. We consider<br />
it most likely that these word breaks are due to a copyist with no knowledge <strong>of</strong> Omagua reproducing another<br />
hanwritten <strong>manuscript</strong> in the time between when the text was last edited and when it was typeset for publication.<br />
113 In Table 3.3, empty cells indicate that there are no attested scribal errors with respect to the particular segment<br />
in question. Recall that grayed out cells indicate that the respective phoneme is not attested in that text.<br />
114 We assume this sequence and its counterpart in the immediately lower cell to be misinterpretations <strong>of</strong> .<br />
115 We assume this to be a misrepresentation <strong>of</strong> , which in all other texts corresponds to /s/.<br />
116 This sequence almost certainly corresponds to , given calligraphic practices <strong>of</strong> the period in which the first<br />
<strong>of</strong> a sequence <strong>of</strong> two s was written as what essentially resembles a cursive .<br />
117 We assume this and its counterpart in the immediately lower cell to be a misrepresentation <strong>of</strong> , which corresponds<br />
to both /e/ and /1/ (see below).<br />
58