Programm 72. StuTS

Event nominals and result nominals with -ing and -ning in Norwegian
Introduction: The article explores Norwegian nominals with the nominalizing suffixes -ing and -ning and whether they are event nominals (EN) or result nominals (RN) (Harley 2009, Lundquist 2008). Most Norwegian verbs can be nominalized with -ing and some can also be nominalized with -ning. Thus, some verbs occur with both suffixes, such as forurensing (‘pollution’) and forurensning (‘pollution’) without a clear semantic difference. According to Språkrådet (The Language Council of Norway) (2017), there is no clear meaning distinction between the two suffixes, while they suggest one can use -ing for an event-oriented meaning and -ning for a result-oriented meaning. The question asked in the article is whether the two suffixes are associated with different semantics and morphosyntax, such that nominals with -ing are instances of ENs, and nominals with -ning instances of RNs. To probe this more systematically, I carried out a corpus study, and found that there is some correlation between suffix choice and type of nominal.
Hypotheses: The question of whether -ing correlates with EN readings and -ning with RN readnings gives rise to three possible hypotheses.
At one extreme, Hypothesis 1 states that there is no correlation between choice of suffix, -ing or -ning, and type of nominal, EN or RN.
At the other extreme, Hypothesis 3 states that there is a strong correlation between suffix choice (-ing or -ning) and type of nominal (EN or RN), such that -ing mainly is used with EN and -ning is mainly used with RN.
Most importantly, a partial correlation is possible, as stated in Hypothesis 2. This can take shape in three distinct ways.
● Subhypothesis 2a: -ing is more often used as EN than RN, whereas -ning is more often used as RN than EN
● Subhypothesis 2b: -ing has a more limited range of use than -ning, such that -ing is more used as EN than RN, whereas -ning is equally often used as EN and RN
● Subhypothesis 2c: -ning has a more limited range of use than -ing, such that -ning is more often used as RN, whereas -ing is used equally often as EN and RN.
Based on the results from the corpus study, I argue in favour of Hypothesis 2 and Subhypothesis 2b.
Annotation criteria: I used four tests as annotation criteria when conducting the corpus study. Diagnostics 1-3 are based on Harley (2009) and Lundquists (2008) characterization of ENs and RNs. Diagnostics 4a and b I propose and argue for myself.
Diagnostic 1 - Count nouns: ENs are mass nouns and RNs count nouns. Therefore, nouns in plural, such as nedlastninger (‘downloads’), indicate RN, since ENs can't pluralize.
Diagnostic 2 - Internal argument: An EN must keep the internal argument of the verb from which it is derived. Thus, internal argument, as in overvåking av beboerne (‘surveillance of the residents’), indicates EN.
Diagnostic 3 - Aspectual modifiers: ENs can be aspectually modified by words like kontinuerlig (‘continuous’) and hyppig (‘frequent’), whereas RNs cannot. I therefore assume that aspectual modifiers indicate ENs.
Diagnostic 4a: Det foregår (‘there is going on’)
As an EN denotes a process or, more generally, an event, and can therefore accompany dynamic verbs such as foregå (‘be going on’) and finne sted (‘take place’). I will assume that an -ing/-ning nominal that is an argument of such a verb is an EN.
Diagnostic 4b: Det finnes (“there exists”)
An RN, on the other hand, denotes a result state or an object, i.e. something that exists. An RN can therefore be an argument of stative verbs such as eksistere (‘exist’) or synonymous verbs, and I will interpret the presence of such a predicate as an indication of an RN.
Method: From the Norwegian Web AS Corpus (NoWaC), I downloaded 100 random samples each of 11 nominals with both -ing and -ning, that is, 2200 samples altogether. I annotated in separate annotation columns every positive result of each test. In separate annotation columns for the type of nominal, samples with positive results on test 2, 3 and/or 4a were assumed to be ENs, and samples with positive results on test 1 and/or 4b were assumed to be RNs. A sample with positive result on both an EN test and an RN test, were annotated as ambiguous. The final step was to compare the percentages of the 1100 -ing samples that were classified as ENs and as RNs, respectively, and repeat the same step for the -ning samples.
Results: Since the samples were naturally occurring corpus examples, 1492 of the 2200 samples did not pass any of the five diagnostics, which test for the presence of certain expressions. I analyzed the remaining 708 samples, which passed one or more of the diagnostics; 339 were with -ing and 369 with -ning, i.e. the data set can be considered balanced. Diagnostics 1 and 2 had the most hits, and the other diagnostics had very few. ENs make up the vast majority of the samples with both -ing and -ning. Nevertheless, there are more ENs with -ing (299) than with -ning (242). With regards to RNs, the gap between the -ing lists and -ning lists is bigger, as there are 108 RNs with -ning and 30 RNs with -ing. There are altogether 29 ambiguous samples. However, the general nature of test 3 gives rise to some false positives with regards to this test. The nominal in (1) passes diagnostic 3, but is more plausibly analyzed as an RN, as the encoded meaning of the proposition is that there exist interpretations of Carmen as a result of how Furman, Paglia and McClary read it.
(1) Eksempler på feministiske lesninger av Carmen er Nelly Furman ( 1988 ) , Examples on feminist reading.PL of Carmen are Nelly Furman (1988), Camille Paglia ( 1994 ) og Susan McClary ( 1992 ) . (8098708, NoWaC) Camille Paglia (1994) and Susan McClary (1992)
‘Examples of feminist readings of Carmen are Nelly Furman (1988) [...] ’ Conclusion: As predicted by Hypothesis 2, the results show a certain pattern with regards to choice of suffix and type of nominal. Furthermore, Hypothesis 2b correctly predicted the use of -ing being mainly limited to EN, whereas -ning is used with both EN and RN. The findings also indicate that the use of internal argument as a distinction between EN and RN (as suggested by Harley (2009) and Lundquist (2008)), yields too many false positives.

Info

Tag: 04.11.2022
Anfangszeit: 08:45
Dauer: 00:30
Raum: Wiwi-Bunker —Room 3035
Track: Corpus Linguistics
Sprache: en

Links:

• iCalendar

Feedback

Uns interessiert Ihre Meinung! Wie fanden Sie diese Veranstaltung?