aan komt vliegen / komt aangevlogenDimensions of Dutch komen ‘come’ + Motion Verb

Jeffrey Pheiff, Forschungszentrum Deutscher Sprachatlas & Lea Schäfer, Heinrich-Heine-Universität Düsseldorf

The PhenomenonAll Germanic languages show periphrastic constructionswith COME and a motion verb. But only in Dutch do wefind morphological variation that may express asemantic distinction (see Beliën 2016: 18; Ebeling 2006:418). Dutch komen ‘come’ can occur with a pastparticiple (1) or an infinitive (2) of a motion verb. Thepresent participle (3) is used to focus on the manner ofthe movement expressed by komen, while INF and PTCPboth highlight the path of the movement itself:

(1) de vogel kwam aangevlogen. PTCP

(2)De vogel kwam aanvliegen. INF

(3)De vogel kwam vliegend. PPR‘The bird came flying.’

Diachronic and Diatopic Variation

The PTCP variant is the older one, while the use of theINF increased in the 13th century (Hirao 1965: 206, 226;Van der Horst 2008: 910). Since the 14th century INF andPTCP have existed in parallel (see the CorpusMiddelnederlands).Based on data from the Meertens Institut’s vragenlijst(sentence no. 52), a DynaSAND question (no. 310), andfindings in dialect dictionaries, we found that the dialectsin the north prefer the INF, while the southeasterndialects favor the PTCP (see Pheiff & Schäfer 2018). Therealso seems to be a transition zone where both stategiesare used in parallel. The picture we get is that the INFhas expanded from north to south over time.

Our Survey

We created an online survey consisting of 10 randomizedpuzzle-judgement tasks, as shown in Figure 4, to test towhat extent the variation we found in the base dialects isstill present in colloquial Dutch and also to test whichvariables influence the use of the INF or the PTCP.Our (preliminary) results are based on 124 completedquestionnaires from the Netherlands with but a handfulfrom the north of Belgium. Due to population density,Holland is better represented than the other regions inour current data set.

Regional Impact

Figure 1: Heatmap INF vs. PTCP

In our data, the INF is the most common form in every region. Only in thesouth (Belgium) does the INF seem to be used to a somewhat lesserextent (see Figure 1). Thus, there is no longer a heavy effect induced bygeographic location in present-day Dutch, contrary to what has beendescribed for the older dialects (Pheiff & Schäfer 2018). Our data supportsCornips’ (2002) findings that there is idiolectal variation w. r. t. this motionverb construction which cannot be explained by any of the “obvious”external facors that we tested (e.g. location, age, gender, education, etc.).

INF vs. PTCP: Semantic ConstraintsAccording to Beliën (2016:30) “the variant with the past participle highlights the end of a process, while the infinitivevariant does not”. This claim could not be corroborated by the outcome of our survey. We only found such an effect withone sentence (ID 04), where half of the informants use the PPR to highlight the special and unusual kind of movement:kwam zij kruipend over de finish ‘she came crawling over the finish line’. By fitting a logistic regression model to our testedvariables (see table 1 below and figure 5 on the right), we found two effects: 1. manner (internal movement) reduces thepresence of INF, and 2. ingressive semantics inhibits the presence of an INF and favors the use of the PTCP. Therefore,we did not find any effect of telicity on the use of the PTCP variant, yet this effect only manifests itself in one direction, i.e.the beginning of the event denoted by the embedded motion verb (see Beliën 2016). However, correlation does not implycausation! Comparing our Aktionsart constraints with the informants’ self-evaluations localizing the movements in the 10sentences (see Figure 2), we see no evidence for a semantic infuence on their choice of INF vs. PTCP (see Figure 3).

Figure 2: Localization of the movement:self-evaluations by 124 informants

Figure 3: Distribution of answers (puzzletest) with INF vs. PTCP in %

ID Lexical Verb PATH/MANNER Aktionsart Speed01 rijden ‘drive, ride’ MANNER egressive ±06 vliegen ‘fly’ MANNER egressive ±07 zwemmen ‘swim’ PATH ingressive ±09 kruipen ‘crawl’ PATH ingressive −10 fietsen ‘drive, ride’ PATH progressive ±11 rennen ‘run’ PATH ingressive +49 vliegen ‘fly’ PATH ingressive ±59 vliegen ‘fly’ PATH ingressive ±05 rennen ‘run’ MANNER ingressive +04 kruipen ‘crawl’ MANNER egressive −

Table 1: The test sentences and their constraints

Particle Split: Syntactic Constraints

Taking a closer look at the syntactic contexts in which the INFs and PTCPs variants can appear, we found no evidence for“particle splits” with PTCP (e.g. ?aan komt gevlogen ‘comes flying’), yet they seem to occur quite naturally with the INF aankomt vliegen. This might be seen as an indication for the grammaticalization of komen towards a modal or even an auxiliaryverb. Such a grammaticalizations cline is typologically frequent and has also been observed with COME + motion verbconstructions in West Germanic languages (see Schäfer submitted). In the survey, we also tested the use of the futureauxiliaries zullen ‘should’ and gaan ‘go’. With these two verbs, particle split is quite common. Our hypothesis is thatparticle split is only possible when komen acts as a modal verb, requiring the infinitive (“first status”, cf. Bech 1955).Therefore, the historical transition from PTCP to INF has almost reached an endpoint in colloquial Dutch, where thetransition was most plausibly triggered by a change of the syntactic function of komen itself rather than specific semanticrestrictions.

Figure 4: Example from the questionnaire (ID 49)

Call:glm(formula = inf == "INF" ~ manner + ingressive + area + verbs, family = binomial, data =

vars)

Deviance Residuals:Min 1Q Median 3Q Max

-2.7070 0.2695 0.4325 0.6272 1.6397

Coefficients:Estimate Std. Error z value Pr(>|z|)

(Intercept) 3.510964 0.814148 4.312 1.61e-05 ***mannerTRUE -1.522116 0.403426 -3.773 0.000161 ***ingressiveTRUE -3.528381 0.641926 -5.497 3.87e-08 ***(and others, not significant)---Signif. codes: 0 ’***’ 0.001 ’**’ 0.01 ’*’ 0.05 ’.’ 0.1 ’ ’ 1

Figure 5: Excerpt from the logistic regression output

ReferencesGunnar Bech. Studien über das deutsche verbum infinitum, volume 1. 1955.Maaike Beliën. Exploring semantic differences in syntactic variation: Dutch komen‘come’ with a past participle or an infinitive. From Variation to Iconicity, page17–32, 2016.Leonie Cornips. Een vreemde eend in het rijtje. over het aspectueelhulpwerkwoord ‘komen’. In verband met Jan Luif, 2002. URLcf.hum.uva.nl/poldernederlands/backup_luif/luif/cornips.htm.Kuzo Hirao. Fügungen des Typs ‘kam gefahren’ im Deutschen. Beiträge zurGeschichte der deutschen Sprache und Literatur, 87:204–226, 2005.Corpus Middelnederlands. Based on the Middelnederlandsch Woordenboek(online). 1250–1500.Jeffrey Pheiff and Lea Schäfer. komt ingefietst/fietsen – On the Diachronic andDiatopic Dimension of Dutch komen COME + Motion Verb. Poster at theSociolinguistics Circle 2018 (Maastricht University: 06.04.2018). 2018. URLhttp://lea-schaefer.de/komt-ingefietst.pdf.Lea Schäfer. ‘kommen’ und Bewegungsverb in westgermanischen Varietäten.Syntax aus Saarbrücker Sicht, 3, submitted.Joop Van der Horst. Geschiedenis van de Nederlandse syntaxis. 2008.*Our thanks go to Michael Cysouw for helping us with the statistics!

Contact– jeffrey.pheiff@staff.uni-marburg.de– lea.schaefer@phil.hhu.de

You can find the survey at: www.soscisurvey.de/beweging/