Íslenskt mál og almenn málfræði - 01.01.2023, Page 164

Now, I would like to ask you a question regarding the quantitative analysis on
the micro-level, when you are looking at the frequency of individual features.
Table 15 shows the most frequent unique features in the data set, the fifteen
most common of which can be ascribed to Icelandic. In your calculations, you do
not lemmatize the words, as you list different word forms as separate entries. For
example, the 1st person pronoun ég only includes the nominative form of the
pronoun. The dative form mér is listed as number 27 and the accusative form
number 43. A similar pattern can be seen for the verb forms er (no. 5 on the list),
var (no. 15), eru (no. 33), and vera (no. 48). Another matter worth mentioning is
the fact that English does not always write compounds as one word and it uses
free-standing articles, the and a, which count as unique features. My question is
as follows:
If we keep in mind the differences between the morphological structure and
spelling conventions of the Icelandic and English languages, isn’t there a risk
that a non-lemmatized frequency table may give a slightly skewed result?

Vanessa Isenmann:
I agree, there is a risk that a non-lemmatized frequency table may give a
slightly skewed result. However, the overall result of the micro-level analy-
sis remains the same, namely that Icelandic features are used to a much big-
ger extent than English or any other features. In the end, this was the main
purpose of the quantitative analysis, to see what linguistic code shows the
most features in the data corpus.

I have a follow-up question regarding the feature which is listed as no. 12 in
Table 15, i/I. You say in your discussion that this feature can occur both in
Icelandic (as an orthographically deviated form of the preposition í) and English
(as the first-person singular personal pronoun). There are also other similar cases
in Tables 15 and 16, for example, the forms of-of and man-man. I assume that all
these words occur in the data in some kind of context which would disambiguate
the meaning. I think that a Facebook friend would immediately know whether
the informant is using the Icelandic verb form man or the English noun man. In
addition, in the case of I, the use of the capital letter would strongly suggest that
the informant is using the English pronoun. So my question is:
Why did you decide to do the frequency analysis without taking the context
into consideration?

Vanessa Isenmann:
That would have been a very good way to approach these features. I cannot
say why I did not do it that way. However, the feature analysis served a very
important purpose that even ambiguous cases such as of-of, man-man or i/I
Helga Hilmisdóttir164