In class, we developed the hypothesis that the inclusion of nouns from a foreign language into a post-monolingual text might in general be more likely than the inclusion of other parts of speech. Our reasoning was that we deem their inclusion into an English sentence and therefore into the English syntax to be relatively easy at least in comparison to verbs, for example, since nouns can be more easily detached from other parts of speech. Verbs are usually closely connected to other parts of the sentence and an adjective would require something it describes. That led us to the perspective of the recipients for whom we argued it would be easier to understand the meaning of nouns in the context of an English sentence than verbs or adjectives precisely because of their close relation to the other parts of speech in a sentence. Furthermore, we discussed that nouns in the context of transcultural exchanges also play an important role, since they can describe important cultural ideas and values – sometimes untranslatable -, so that with nouns, we can also think of the politics of translation and especially of untranslatable expressions. For example, we could ask what effects a foreignizing or domesticating translation strategy with regard to these expressions could have – also in terms of global power hierarchies that were often accompanied by an imposition of a certain language by the colonial power.
When I took another look at the sentences I had decided to annotate, which were then added to our seminar’s corpus, I remembered that I had tried to find sentences with the greatest variance possible in what kinds of non-English words were included, how many non-English words in a row, and in what placement of the sentence the non-English words were included. Therefore, within my example sentences, there are not only non-English nouns, but (among others) also non-English adjectives. One of them being the following:
‘ A kawaida job like this ? ’ he
False False True False False False False False False
PUNCT DET NOUN NOUN ADP PRON PUNCT PUNCT PRON
–
Had whispered as he lowered her .
False False False False False False False
AUX VERB SCONJ PRON VERB PRON PUNCT
(False = English word, True = non-English word)
In this case, I would say that even though a non-English adjective describes an English noun, the meaning can be understood, if not precisely, then at least roughly, thanks to the content of the surrounding sentences. While in some other instances in the novel Dust by Owuor, the non-English words are immediately followed by their English equivalents, this is not the case with the Kiswahili word ‘kawaida’ (‘normal’, ‘commonplace’, ‘everyday’). This shows that in the practice of non-monolingual Anglophone writing, foreign adjectives are also used without explaining them in English. Unfortunately, I am not in a position to say whether there is actually a tendency towards a part of speech in the inclusion of non-English words, firstly because the corpus is mainly designed with a view to the greatest possible variance and secondly because the program failed to annotate the parts of speech for the non-English words because it can only ever be fed with one language model, so that although the English words are correctly annotated, the non-English words were in most cases labeled as nouns or proper nouns, no matter their actual category.
Something I found very enlightening in this regard was the following sentence:
The German continues , ‚ A trival pis
False False False False False False False True
DET ADJ VERB PUNCT PUNCT DET NOUN NOUN
–
festival vizin ze allegorical oasis … ‚
False True True False False False False
NOUN NOUN PROPN ADJ NOUN PUNCT PUNCT
On the one hand, this sentence is interesting because it contains the expression ‘the German’ in which ‘German’ functions as a noun, not an adjective, as which the machine mistakenly labeled it. On the other hand, readers can easily understand that the direct speech is in fact composed in the English language, the words are just written differently to indicate the strong German accent of the speaker. We can understand that what the man is saying would look like this in standard English orthography: ‘A tribal peace festival within the allegorical oasis …’. In contrast, the program was not able to recognize the words as English words and consequently labeled those words as either nouns or proper nouns. From this I conclude that the machine strictly works on the basis of patterns and probabilities which leads to problems not only with foreign words but also with English words that differ in their orthography or their usage in a sentence. The case of the ‘German’ as a noun that the machine did not recognize as such is especially surprising to me, since the word is even located at the beginning of a sentence and it is accompanied by a determiner which the program recognized. The conclusion that it must be a noun then, is actually very obvious. A question I asked myself was why the program keeps labeling those words it does not recognize as nouns and proper nouns. Is it because nouns play such an important role in the English language? Is it because there really is a tendency toward the inclusion of nouns as foreign words and the program has this information? (I doubt that.) From, the example sentences I was able to conclude that the labeling does not depend on the position of the non-English words. No matter where they are inserted, they will be labeled as a noun or proper noun. Whether the word will be labeled as a proper noun or a noun also seems to be completely arbitrary. I cannot find any reason why ‘vizin’ would be a noun but ‘ze’ a proper noun.