Content uploaded by Daniel Van Olmen
Author content
All content in this area was uploaded by Daniel Van Olmen on Jul 10, 2023
Content may be subject to copyright.
1
What we can learn from using a visual questionnaire to investigate
Dutch and Afrikaans impersonal strategies
Abstract
The topic of impersonalisation has received a lot of attention in the literature, but the focus has
mostly been on a limited number of strategies, such as the use of personal and indefinite pro-
nouns and passive constructions. Impersonal strategies have thus far been examined using: (i)
grammars, (ii) corpora; and (iii) language-based questionnaires. These methods suffer from
several shortcomings if one wants to study the range of impersonal strategies. The present ar-
ticle aims to argue for a new way of investigating impersonal strategies that complements the
other approaches, by reporting on the results of a visual questionnaire. More precisely, it dis-
cusses a visual questionnaire completed by speakers of Dutch and Afrikaans to determine
whether this method is a satisfactory way of studying impersonal strategies and to also examine
and compare the impersonal strategies of the two languages.
2
1 Contextualisation
Impersonal strategies
1
are strategies in which the first argument is not grammatically expressed
or performs a pleonastic function only
2
. It is, in truth, semantically empty, whether marked or
unmarked (compare Siewierska, 2008). In other words, impersonal strategies can be defined as
strategies that contain no referential first argument (Malchukov & Siewierska 2015:20). Maz-
zitelli (2019:32) refer to these kinds of strategy as “agent-defocusing constructions”. From the
literature (e.g. Gast & van der Auwera 2013; Siewierska & Papastathi 2011), we know that
there are two main types of context that such strategies can be used for. Universal contexts,
such as one only lives once, involve a generic first argument that can be paraphrased as ‘eve-
ryone, anyone’. Existential ones, like my car has been stolen, have a specific but unidentified
(group of) individual(s) as their first argument, which can be paraphrased as ‘someone, some
people’. Further distinctions within these types can be made but they will be discussed in more
detail in Section 2.
1
The first sections of this article show a strong similarity with the contextualisations and
theoretical descriptions in [Anonymised] 2021a; 2021a. This overlap is inevitable, however,
since the present publication forms part of the same research project, reports the results of
the questionnaire discussed in [Anonymised] 2021a and also partially involves the data dis-
cussed in [Anonymised] 2021b.
2
For the purpose of this article, we adopt a broad definition of "impersonal strategies", in-
cluding any strategy employed by a language user that lacks a referential subject. Thus, we
classify any strategy used by a speaker to avoid explicitly naming or referring to an agent
as an impersonal strategy in this article.
3
Much recent research has been concerned with impersonal strategies (e.g. Kitagawa & Lehrer
1990; Luukka & Markkanen 1997; Egerland 2003; Hoekstra 2010; Primus 2011; Siewierska
& Papastathi 2011; Gast & van der Auwera 2013; Kirsten 2016; Van Olmen & Breed 2018a,
2018b; Mazzitelli 2019; Van Olmen et al. 2019; Schlund 2018, 2020; [Anonymised] 2021a;
Prenner & Bunčić 2021; Bauer 2021; Groenen 2021). It has, however, mainly focused on a
fairly limited number of strategies such as pronominal ones (e.g. one) and passives (e.g. has
been stolen). Few studies have tried to examine the range of impersonal strategies that are
available to speakers and/or determine which of the strategies they prefer in different imper-
sonal contexts. Two exceptions to this gap in the research are Siewierska's (2008) investigation
into pronominal versus verbal impersonal strategies, although her work is primarily based on
grammatical descriptions and input from a small set of informants, and Bauer's (2021) investi-
gation of the impersonal strategies in six Slavic languages, based on a parallel corpus. Bauer
identifies no less than eighteen distinguishable strategies for impersonalisation in these six
Slavic languages. However, to the best of our knowledge, there has been no systematic analysis
of the variety of impersonal strategies in extensive West-Germanic language data.
Impersonal strategies have thus far been examined using: (i) grammars, often in combination
with first language speaker judgments (e.g. Siewierska 2011; Gast & van der Auwera 2013);
(ii) corpora (e.g. Marin-Arrese et al. 2001; Primus 2011; Coussé & van der Auwera 2012;
Schlund, 2020; Bauer 2021); and (iii) language-based questionnaires (e.g. Siewierska 2008;
Garcia et al. 2018, Prenner & Bunčić 2021). These methods suffer from a number of shortcom-
ings if one wants to study the range of impersonal strategies in a language (see [Anonymised]
2021a). They are, for one, generally deductive in that they take a predetermined set of strategies
as their point of departure (e.g. a questionnaire then asking for acceptability judgments about
them). Admittedly, parallel corpora do not have this drawback: a trigger in the source language
(e.g. German man ‘one/they’ in Gast 2015 and Bauer 2021) may be rendered in the target
4
language(s) in a non-predetermined variety of ways. Still, results may be affected by interfer-
ence from the source language and/or the translation process (see Schlund 2020:56, Bauer
2021:153-156). Moreover, parallel corpus studies share the problem with corpus research in
particular that some impersonal contexts are quite rare
3
in text collections (e.g. those tied to the
here and now of the situation; see Section 2) – making it difficult to find out which strategies
are used in them. Questionnaires in turn often have the disadvantage that they limit the replies
that participants can give (e.g. a completion task where the subject slot in … has/have stolen
my car allows for someone and they but not for a passive).
The present article aims to argue for a new way of investigating impersonal strategies that
would make it possible to identify their variety in different impersonal contexts and demon-
strate which ones are preferred in these contexts and that thus complements the other apprao-
ches. More precisely, it discusses a visual questionnaire completed by speakers of Dutch and
Afrikaans and thus seeks to (i) determine whether this method is a satisfactory way of studying
impersonal strategies and (ii) examine and compare the impersonal strategies of Dutch and
3
One of the reviewers of our article emphasised that the issues and challenges related to cor-
pus investigations that we discussed are primarily relevant to the current state of available
corpora. However, as spoken language face-to-face interaction corpora become increasingly
accessible, many of these problems may be mitigated. Such corpora may contain examples
of language use that are not frequently encountered in written language, thereby filling gaps
in existing corpora. Nevertheless, for certain under-resourced languages like Afrikaans, de-
veloping these corpora may take more time. As a consequence, researchers investigating
such languages may still encounter some of the gaps and challenges that are typical of tra-
ditional corpus investigation.
5
Afrikaans. The rest of the article is structured as follows. Section 2 briefly introduces the im-
personal contexts distinguished in the questionnaire. In Section 3, we explain the method and
the design of the visual questionnaire in more detail and, in Section 4, we discuss its results for
Dutch and Afrikaans. Section 5, finally, is the conclusion.
2 Impersonal contexts
The visual questionnaire distinguishes twelve different impersonal contexts. They are based on
Van Olmen & Breed’s (2018b) semantic map, which itself combines criteria that feature in
Siewierska & Papastathi’s (2011) and Gast & van der Auwera’s (2013) semantic maps. As can
be expected from semantic maps, all distinctions are motivated by cross-linguistic variation
(e.g. the context in they say that the house is haunted differs from that in they have stolen my
car since some languages accept the third person plural in the one but not the other). An in-
depth discussion of this evidence is beyond the scope of the present article but can be found in
the aforementioned sources.
The twelve contexts, which are exemplified below, can be distinguished from each other based
on seven parameters: (i) quantification; (ii) perspective; (iii) veridicality; (iv) modality; (v)
(un)knownness; (vi) number; and (vii) speech act. These parameters will now be discussed one
by one, with reference to the examples.
Impersonal context
English example
1
UNI-INT-NVER-
NMOD
(1) But what happens when you get laid off?
2
UNI-INT-NVER-MOD
(2) One should not be ungrateful.
3
UNI-INT-VER
(3) You only live once.
4
UNI-EXT
(4) In Belgium, they are proud of their fries.
6
5
EXI-COR
(5) Tunnel users should be aware of the various speed cam-
eras that they have installed in the tunnel.
6
EXI-VAG-PL
(6) A leopard has been spotted in several places in the vil-
lage.
7
EXI-VAG-NN
(7) Someone has stolen his clothes.
8
EXI-INF-PL
(8) If we look at the archaeological evidence, we see that
they built a settlement here.
9
EXI-INF-NN
(9) He points to trampled leaves. Someone has already
walked here.
10
EXI-SPE-PL
(10) We heard them coming in from the back door and the
front door at the same time.
11
EXI-SPE-NN
(11) I have to go. Someone is waiting for me in the lobby.
12
SAV
(12) They say there is evidence that corporal punishment cre-
ates a culture of violence.
Table 1. Twelve distinguishable impersonal contexts
Quantification involves the distinction introduced in Section 1 between universal (UNI) and
existential (EXI) uses. The former apply to everyone contextually relevant, like in (1) to (4),
while the latter concern one or more particular but unidentified individuals, like in (5) to (12).
Perspective is a parameter that differentiates universal uses and centres, in essence, around the
(non-)inclusion of the speaker and the addressee in the set. A universal use that applies to
speaker and addressee too, like (1) to (3), has an internal (INT) perspective. One with an exter-
nal (EXT) perspective excludes them, as in (4). This sentence is a statement about all people
in Belgium but the speaker and the addressee clearly do not belong to this group (cf. we/you
are proud of our/your fries in Belgium).
Veridicality distinguishes universal-internal uses from each other and involves the presentation
of the state of affairs as real or not. Example (3) is veridical (VER) since only living once is
given as a fact (of life). Examples (1) and (2), however, are non-veridical (NVER): getting laid
off is presented as being in the realm of the hypothetical and being ungrateful as being in the
realm of the undesirable.
7
Modality further differentiates non-veridical universal-internal uses. If non-veridicality is ex-
pressed by some overt modal element, the use is modal (MOD). The auxiliary should in (2) is
a case in point. If non-veridicality is conveyed by other means (e.g. the interrogative nature of
the sentence), the use is non-modal (NMOD). In (1), for instance, the non-veridicality comes
not from a modal element but from the conditional character of the subordinate if-clause.
(Un)knownness is a parameter that pertains to existential uses. It has to do with the amount and
type of information that is available about the particular but unidentified (set of) individual(s).
Four distinctions are made here. First, in partially known or so-called corporate (COR) uses, it
is relatively clear in a way from the state of affairs itself who is responsible for it, even if they
are not explicitly named. Typically, they are some kind of institutional entity – hence, the term
“corporate”. In (5), for example, the state of affairs of installing speed cameras is something
that can only really by realised by the police and/or the agency in charge of road signs and the
like. Second, in vague (VAG) uses such as (6) and (7), the speaker really knows about the event
being described but is not able or willing to identify the particular person or people responsible
for it. Third, in inferred (INF) uses like (8) and (9), the speaker does not have any actual direct
knowledge of the event. They deduce its existence from other evidence available to them and
then also assume that some unknown person or people must be behind it. In (8), for instance,
the speaker infers from the archaeological data that there was at some point a settlement where
they are and thus also that some community building it must have existed. Fourth, SPE uses
refer to a specific point in time (see Siewierska & Papastathi, 2011:582). In examples such as
(10) and (11), the speaker is in the same place and time as the individual(s) realising some
event there and then and may thus have strong suspicions about who they are but is still not
explicitly identifying them. In (11), for example, there is a person currently waiting for the
speaker, who probably knows who this individual is but chooses not to name them in their
utterance.
8
Number as a parameter intersects with the vague, inferred and specific existential uses (not
with the corporate ones, though, since they are inherently plural, involving entities like the
government, the hospital and so forth). In each of these three contexts, we can have a state of
affairs that necessarily involves more than one person – like in (6), (8) and (10). These exam-
ples are, in other words, plural (PL). We can, however, also have a state of affairs in each
context that may be realised by one or more than one individual – like in (7), (9) and (11).
These examples are thus number-neutral (NN).
Speech act verb (SAV), lastly, sets apart one particular impersonal context from all others. It
involves the presence of a speech act verb that fulfils an evidential function in the sentence,
like say in (12). The speaker here attributes a claim to an unspecified set of people.
3 Methodology
3.1 Questionnaire
As discussed in Section 1, current methods may not be entirely suitable for identifying the full
range of impersonal strategies in a language, or for determining the preferred strategies in dif-
ferent impersonal contexts. Let us illustrate this point here in more depth, with a look at Van
Olmen & Breed’s (2018a, 2018b) methodology. They adopted a “double questionnaire-based
approach” to study impersonal strategies in West Germanic. A first group of first language
speakers of English, Dutch and Afrikaans were given an acceptability judgment questionnaire
9
(see Figure 1), and a second group a completion task questionnaire (see Figure 2).
Figure 1. Acceptability judgment stimulus for UNI-INT-NVER-MOD (Van Olmen & Breed
2018b)
Figure 2. Completion task stimulus for UNI-INT-NVER-MOD (Van Olmen & Breed 2018b)
Both questionnaires contained twenty-four stimuli (in an arbitrary order), two for each of the
twelve impersonal uses distinguished in Section 1. Figure 1 and Figure 2, for instance, present
one of the two universal-internal non-veridical-modal stimuli. The first questionnaire invited
the respondents to assess the acceptability of a number of impersonal strategies as a way to
complete the scenario described above, on a five-point scale where one stands for very unac-
ceptable and five for very acceptable. The list of impersonal strategies to be judged only con-
tained a limited set of pronominal ones, however (see one, you and they in Figure 1). As a
result, we do not know how their acceptability would compare to other potential “solutions”
like people cannot learn a language in six weeks. The second questionnaire was intended to
10
address this issue and asked respondents to fill in the subject slot of the final sentence of each
stimulus themselves. As Figure 2 shows, this approach did allow respondents to use not only
one and you but also people for UNI-INT-NVER-MOD contexts. The sentence with the blank
excluded a whole range of other conceivable answers, though – like negative indefinite nobody
can learn a language in six weeks, non-finite learning a language in six weeks is impossible or
passive a language cannot be learnt in six weeks. In fact, one respondent seemed to feel so
strongly about the passive for one of the stimuli in the completion task that they simply ignored
the structure of the sentence to be completed.
Despite the problems with the above methods, a questionnaire-based approach still has consid-
erable promise for a study of the variety of impersonal strategies in a language. The reason is
that, unlike corpus research, for example, it enables us to examine, through targeted stimuli,
impersonal contexts that do not occur very often in usage. The method adopted in the present
article therefore sticks with presenting respondents with a scenario for all twelve impersonal
uses that prompts them to complete it with an impersonal strategy. The scenarios are, however,
given not as descriptions, like in Figure 1 and Figure 2, but as visual representations, like in
Figure 3.
4
4
The stimuli were created through a collaborative, interdisciplinary project based in linguistic
theory about impersonalisation and visual communication theory about wordless visual nar-
ratives (e.g. Nodelman 1990; Horwat 2018:176; Arif & Hashim 2008:121). The collabora-
tors consisted of two linguists, one lecturer in graphic design and fifteen students in graphic
design. The questionnaire was developed as a practice-based research project, with the aim
of training students in graphic design to navigate the constraints and interactions of a client
assignment. The linguists were the client and their commission brief stated that the students,
11
Figure 3. Visual questionnaire stimulus for EXI-INF-PL (‘Give an appropriate utterance for
the speech bubble. Please use a form of the verb “play” (e.g. play, plays, played, playing) and
the word “football” in your answer.’)
To obtain as many useful answers as possible, we instructed the participants to provide a spe-
cific type of response on the opening page of the questionnaire. The instructions explicitly
stated that their answer should be focused on people in general or individuals that they do not
under their lecturer’s guidance, had to produce visual representations, of the various imper-
sonal uses, that could function as “visual questions or directions” for a questionnaire. See
[Anonymised] 2021a for more information and an evaluation of the method.
12
know or cannot identify. Therefore, responses that were only related to the participant them-
selves, or to a specific person or group, were discouraged.
In addition to these instructions, we also incorporated visual cues and prompts to further guide
the respondents towards providing useful answers. These visual stimuli were accompanied by
indications of the type of response that was expected. They mostly constitute requests that
respondents use a specific verb or sometimes also a specific adverb or noun to complete a
scenario, as can be seen in Figure 3 with spelen ‘play’ and voetbal ‘football’ and in the rest of
the questionnaires, which are accessible online
5
. Even with such indications, however, the
stimuli for the speech act verb use (e.g. ‘it is said that this house is haunted’) have proven to be
inadequate, with too many irrelevant answers, and are therefore not discussed in the remainder
of this article. For the other eleven uses, examples (13) to (34) illustrate suitable answers actu-
ally offered by our respondents.
Impersonal
context
Suitable examples from answers of
Dutch visual questionnaire
Suitable examples from answers of
Afrikaans visual questionnaire
1
EXI-COR
(13) Kijk schat, er wordt aan de
weg gewerkt.
‘Look darling, the road is be-
ing worked on.’
(14) Hulle werk alweer aan die pad.
‘They are working on the road
again.’
2
EXI-INF-NN
(15) Kijk nou wat voor slipspoor ze
hebben gemaakt.
‘Just look at the sort of skid
mark that they have made.’
(16) Hulle het natuurlik gister hier
gedrink en nie skoon gemaak
nie.
‘They were obviously drinking
here yesterday and didn’t clean
up.’
5
An anonymised version of the Dutch visual questionnaire:
https://forms.gle/TgUQzYfDFnDMV8QE7; an anonymised version of the Afrikaans visual
questionnaire https://forms.gle/qQUkx8gxSm8u6GbT6
13
3
EXI-INF-PL6
(17) Iemand heeft gisteravond een
vuurtje gestookt.
‘Someone lit a fire here yes-
terday evening.’
(18) Lyk of iemand gister hier sokker
gespeel het.
‘Looks like someone played
football here yesterday.'
4
EXI-SPE-NN
(19) Iemand probeert in te breken!
‘Someone is trying to break
in!’
(20) O gats! Iemand het my gehoor!
‘O no! Someone heard me!’
5
EXI-SPE-PL
(21) Ze hebben de hele taart
opgegeten!
‘They have eaten the whole
tart.’
(22) Iemand het sowaar al die koek
opgeëet!
‘Someone has, can you believe
it, eaten the whole cake.’
6
EXI-VAG-NN
(23) Jouw zoon wordt gepest op
school!
‘Your son is being bullied at
school.’
(24) Iemand het al die appels gepluk
terwyl jy geslaap het!
‘Someone picked all the apples
while you were asleep.’
7
EXI-VAG-PL
(25) Alle computers werden
gestolen.
‘All computers were stolen.’
(26) Hulle het terreuraanvalle op ’n
paar plekke in die stad
uitgevoer.
‘They have carried out terror-
ists attacks at a couple of places
in the city.’
8
UNI-EXT
(27) In de woestijn rijden ze op
kamelen.
‘In the desert, they ride on
camels.’
(28) Mense in China eet baie rys.
‘People in China eat a lot of
rice.’
9
UNI-INT-
NVER-MOD
(29) Als je gaat rijden moet je een
autogordel omdoen.
‘if you are going to drive, you
have to put on a seatbelt.’
(30) ’n Mens moet altyd jou
sitplekgordel vasmaak as jy
motor bestuur.
6
One of the reviewers has raised a concern that the EXI-SPE-PL context in our questionnaire
may be perceived by participants as EXI-SPE-NN, as starting a fire can be a task performed
by a single individual. While we acknowledge that this is a possibility, we would like to
emphasise that the stimulus includes multiple seats around the fire, which indicates a group
activity. Furthermore, it is possible that the potential for misinterpretation exists for all vis-
ual questions in the questionnaire. However, we believe that one of the strengths of our
questionnaire design is that each illustration is carefully crafted to visually depict the spe-
cific criteria of the question, while also using the principles of visual narrative theory to
eliminate potential misinterpretations.
14
‘One should always fasten
one’s seatbelt when one
drives.’
10
UNI-INT-
NVER-NMOD
(31) Kun je doodgaan van een
bijensteek?
‘Can you die from a bee
sting?’
(32) Wat gebeur as mens dié rooi
knoppie druk?
‘What happens if one presses
that red button?’
11
UNI-INT-VER
(33) In de winter krijgen we het
koud.
‘In winter, we get cold.’
(34) Almal kry koud in die winter.
‘Everyone gets cold in winter.’
Table 2. Examples of responses for each impersonal context from the Dutch and Afrikaans
questionnaires.
3.2 Dataset
The visual questions in the Dutch and Afrikaans questionnaires were identical, but the textual
explanation for each question was translated into the language of the questionnaire. The ques-
tionnaires were distributed via social media. Participation was anonymous and voluntary. A
total of 83 first language speakers of Dutch ended up completing the questionnaire for their
language. For the Afrikaans questionnaire, we managed to get no less than 454 first language
speakers to fill it out. Because of this discrepancy between the two languages and the amount
of data for the latter in particular (more than 10,000 data points), we analysed all of the data
for Dutch but only the responses for one visual stimulus for each impersonal use for Afrikaans.
The Afrikaans data has been reported on already, in [Anonymised] 2021b, but is included here
nonetheless as a basis for comparison with the results for Dutch. An overview of the data is
given in Table 3. It presents, for each impersonal context (see the leftmost column): (i) the
number of stimuli taken into consideration here (e.g. for Afrikaans, always just one); (ii) the
number of irrelevant answers; (iii) the number of relevant answers; and (iv) the total number
of answers.
15
Context
Dutch dataset
Afrikaans dataset
Amount of questions
included
Usable answer per
context
Not usable answers per
context
Total
Amount of questions
included
Usable answer per
context
Not usable answers per
context
Total
SAV
Not included in analysis
Not included in the analysis
EXI-COR
2
122
42
164
1
362
77
439
EXI-INF-NN
2
122
120
242
1
368
78
449
EXI-INF-PL
2
94
67
161
1
220
227
448
EXI-SPE-NN
2
149
15
164
1
187
224
411
EXI-SPE-PL
2
116
48
164
1
245
206
451
EXI-VAG-NN
2
141
25
166
1
343
107
450
EXI-VAG-PL
2
149
16
165
1
424
25
449
UNI-EXT
3
168
79
247
1
288
165
453
UNI-INT-NVER-MOD
4
280
49
329
1
360
92
452
UNI-INT-NEVER-NMOD
4
281
11
292
1
382
68
450
UNI-INT-VER
3
206
45
251
1
119
335
454
TOTAL
1828
517
2345
1
3298
1604
4906
Table 3. Overview of the visual questionnaire data for Dutch and Afrikaans
A few comments are in order here. First, the totals do not always add up to the same number,
since not all respondents completed all questions and some impersonal uses were tested more
than two times for Dutch. The higher numbers of stimuli for certain uses are an artifact of the
questionnaire development (see footnote 4). Each designer was asked to create a stimulus for
at least one existential context – with eight to choose from – and at least one universal context
– with just four to choose from. As a result, more stimuli were produced for universal uses.
Second, the so-called irrelevant answers include not only those that cannot be considered as
impersonal in any way but also those that may be impersonal but do not actually fit the imper-
sonal use that the stimulus sought to test. An example would be an Afrikaans clause with mens
‘one’, which is exclusively universal-internal (see Van Olmen et al. 2019), for a stimulus
16
looking for existential impersonal strategies). Third, for the relevant answers, we did not ex-
clude those possibly ambiguous between an impersonal and a non-impersonal reading. Dutch
clauses with the pronouns ze ‘they’ or je ‘you’, for instance, could be taken to refer to a known
group of people or an addressee respectively but they can have an impersonal interpretation
too and are thus taken into account as relevant here. Fourth, given the open-ended nature of the
questionnaires and the complexity of the domain under investigation, it is not unsurprising that
there are so many irrelevant answers. Still, we have a sufficient amount of relevant ones for
both languages and all impersonal contexts to study their variety in impersonal strategies in
Section 4. Fifth, and finally, the quantitative differences in the answers between Dutch and
Afrikaans pose no significant problem, since we do not explicitly seek to compare the frequen-
cies of impersonal strategies in the two languages.
4 Results
In this section, we first offer an overall picture of the findings of our visual questionnaires for
Dutch and Afrikaans, focusing on the most frequent types of impersonal strategies (Section
4.1). We then move on to a discussion of the other strategies that can be distinguished in the
data (Section 4.2). We end with a survey of all distinct impersonal uses and the strategies that
are used for them (Section 4.3).
4.1 Overall results and main impersonal strategies
Table 4 and and Error! Reference source not found. provide an overview of our findings for
Dutch and Afrikaans by singling out the most common impersonal strategies in the data. They
are: personal pronouns (PN.PRS), indefinite pronouns (PN.INDF), nouns meaning ‘human be-
ing’ or pronouns originating from such nouns (NPHUMAN) and passives (PASS) (the remaining
relevant answers are included as ‘other’). The term "main strategies" is appropriate for referring
to these particular techniques, as they are frequently employed by participants and are also
17
widely recognised as the central impersonal strategies in existing linguistic literature. In addi-
tion to these main strategies, our results also presented other impersonal strategies that have
not been previously documented in existing linguistic literature. For the purposes of Table 4,
we will refer to these novel strategies as "other strategies”, but we will provide a detailed dis-
cussion of each of these strategies in the subsequent section. The table makes further distinc-
tions for the first three categories (see the second column) and gives the raw (#) and relative
(%) frequencies of all (sub)categories. Figure 4 presents the proportions of the main categories
in graph form.
Strategy
Dutch
Afrikaans
#
%
#
%
PN.PRS
‘I’
96
5%
306
9%
‘you’
387
21%
127
4%
‘we’
78
4%
16
0%
‘you all’
0
0%
3
0%
‘they’
232
13%
554
17%
Total
793
43%
1006
31%
PN.INDF
iemand ‘someone’
158
9%
473
14%
allemaal (D) almal (A) ‘all’
13
1%
40
1%
iedereen (D), elkeen (A) ‘each one’
85
5%
2
0%
Total
256
14%
515
16%
NPHUMAN
men (D) ‘one/they’, ('n) mens (A) ‘one’
49
3%
123
4%
(de) mensen (D), (die) mense (A) ‘(the)
people’
83
5%
189
6%
Total
132
7%
312
9%
PASS
413
23%
955
29%
Other impersonal strategies
234
13%
510
15%
Total
1828
100%
3298
100%
Table 4. Impersonal strategies in Dutch and Afrikaans
18
Figure 4 – Main impersonal strategies in Dutch and Afrikaans
A first thing to observe is that personal pronouns are the most frequent type of impersonal
strategy in both Dutch (43%) and Afrikaans (31%), as in (35) to (38) and (39) to (43) respec-
tively.
(35) Du Wat zou gebeuren als ik een kauwgomballon té groot zal opblazen?
‘What would happen if I blow up a chewing gum balloon too big?’
(36) Du Je mag hier roken blijkbaar.
‘Apparently, you are allowed to smoke here.’
(37) Du We krijgen het koud in de winter omdat de zon niet schijnt.
‘We get cold in winter because the sun doesn’t shine.’
(38) Du Ze hebben alle appels geplukt terwijl je sliep.
‘They picked all the apples while you were asleep.’
(39) Afr Ek maak altyd my sitplekgordel vas
‘I always fastern my seatbelt.’
(40) Afr Wat gebeur as jy die knoppie druk?
‘What happens if you press the button?’
(41) Afr In die winter kry ons koud.
‘In winter, we get cold.’
(42) Afr Maak seker dat julle sitplekgordels vasgemaak is wanneer julle ry.
‘Make sure that your seatbelts are fastened when you drive.’
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Dutch Afrikaans
PN.PRS PN.INDF NP PASS Other
19
(43) Afri Hulle werk alweer aan die pad.
‘They are working on the road again.’
Admittedly, their proportions may be somewhat inflated: strictly speaking, we do not know
whether respondents intended ‘you’, ‘we’ or ‘they’ as impersonal or as referring to particular
people. This ambiguity
7
is especially pertinent for ‘I’. The first person singular is known to be
able to function impersonally (e.g. Zobel 2016) but it is not unlikely that respondents actually
used it to refer to the speaker
8
. Such qualifications notwithstanding, it is evident from the data
7
Although it was not always possible to determine whether a respondent in these cases in-
tended the answers as impersonal, we counted all cases of ambiguity (in other words the
answer can be considered personal or impersonal) as impersonal. The reason for this is that
the ambiguity in itself is an indication that the particular strategy may be a strategy to ex-
press impersonal meaning. Secondly, the respondents - as already mentioned - received a
general instruction at the beginning of the questionnaire which says that their impersonal
answers should be focused on people in general or individuals that they do not know or
cannot identify.
8
One of our reviewers rightly pointed out that the fact that the questionnaires can also yield
ambiguous answers means that, in this respect, this method still has the same limitation as
corpus studies. This is, of course, true. However, a questionnaire approach has two ad-
vantages over corpus investigations in this case. First, the questionnaire designer can include
instructions in the questionnaire that ask the respondent in advance not to offer ambiguous
answers (as was also the case with this questionnaire). A second potential advantage of a
questionnaire approach (one that we did not utilise in the design of this questionnaire) is to
ask respondents follow-up questions as part of the questionnaire design - to clarify the
20
that Dutch and Afrikaans are very partial to employing personal pronouns for impersonal con-
texts. This finding of our visual questionnaire can be seen as an independent justification and
perhaps also as an explanation for much of the research’s focus on the impersonal use of pro-
nouns such as ‘you’ and ‘they’ in Dutch and Afrikaans (e.g. Van Olmen & Breed 2018a, 2018b;
Groenen 2021).
The passive comes out as a very common impersonal strategy too, in Dutch (23%) and in Af-
rikaans (29%). For respective examples, consider (44) and (45).
(44) Du In de woestijn wordt er op kamelen gereden
‘In the desert, camels are ridden on.’
(45) Afr Terwyl jy geslaap het is al die appels gepluk.
‘While you were asleep, all the apples were picked.’
The passive might even prove to be more frequent than personal pronouns
9
, if we were able to
identify and discard those cases in which ‘they’ and the like were not meant as impersonal. In
view of its rate of occurrence in our data, it is quite interesting that the passive has received
comparatively less attention in the literature on impersonal strategies than personal pronouns
meaning of their answers. Therefore, if the questionnaire were not a closed/anonymous
questionnaire process, we would ask all respondents who gave personal pronouns as an an-
swer if they intended an impersonal or personal reference.
9
One of our reviewers noted that this is an intriguing finding from the questionnaire since
passive constructions are often considered to be only a feature of formal writing. The results
demonstrate that passives are also frequently used in informal everyday language (cf. Pren-
ner & Bunčić 2021: 217 on the -no/-to impersonal construction in Polish).
21
(but see Breed & Van Olmen 2021 on the impersonal passive in Dutch and Afrikaans, and
Primus 2011 on the impersonal passive in German and Dutch).
Indefinite pronouns make up the third largest category in Dutch (14%) and Afrikaans (16%).
The questionnaire answers in (46) to (50) are cases in point.
(46) Du Iemand heeft alle computers gestolen.
‘Someone stole all computers.’
(47) Du Iedereen moet een autogordel omdoen, klein en groot.
‘Everyone should put on a seatbelt, whether they are young or old.’
(48) Afr Ek dink iemand het my gehoor
‘I think someone heard me.’
(49) Afr Almal moet hulle sitplekgordel vasmaak.
‘All should fasten their seatbelts.’
(50) Afr Elkeen moet hul sitplekgordel vasmaak.
‘Everyone should fastern their seatbelts.’
Nouns meaning ‘human being’ and pronouns deriving from such nouns, finally, account for
7% of the Dutch answers and 9% of the Afrikaans ones. Some examples are provided in (51)
to (54).
(51) Du In Frankrijk eet men croissants.
‘In France, they eat croissants.’
(52) Du Als mensen uit hun mond stinken, poetsen ze hun tanden en sommigen ook hun tong.
‘If people have a smelly mourth, they brush their teeth and some also their tongues.’
(53) Afr In die winter kry mens koud.
‘In winter, one gets cold.’
(54) Afr Mense in China eet baie rys.
‘People in China eat a lot of rice.’
The frequency of Afrikaans (’n) mens ‘one’, compared to that of jy ‘you’, in Table 4 is note-
worthy. Previous research, in particular Van Olmen & Breed (2018a), suggests that speakers
of Afrikaans strongly prefer the NPHUMAN option to the second person singular. In our visual
22
questionnaire, however, (’n) mens essentially occurs as often (123 times, 3.73%) as jy (127
times, 3.85%). The specific stimuli may have played a role here and the choice between the
two clearly deserves to be studied in more detail (Dutch behaves as expected in Table 3 when
it comes to men ‘one/they’ versus je ‘you’(and ze ‘they’): the NPHUMAN is much less frequent,
with just 49 cases, than the second person singular, with 329 instances(, and the third person
plural, with 221 hits)).
After the above overview of the most frequent types of impersonal strategies in Dutch and
Afrikaans, which are also the most well-established ones in the literature, we now turn to the
answers labelled as ‘other’ in Table 2.
4.2 Other strategies
Table 5 summarises the ‘other’ strategies. Examples and a description of each one will be pre-
sented in the following subsections. The frequencies of the various impersonal strategies used
in different contexts are presented and discussed in detail in Section 5.3.
Strategy
Dutch
Afrikaans
Total
% of total
Total
% of total
Specified NP
99
42,49%
221
43,33%
Imperative
17
7,30%
124
24,31%
Relativisation
7
3,00%
81
15,88%
Infinitive
28
12,02%
38
7,45%
Nominalisation
10
4,29%
12
2,35%
Agentive NP
31
13,30%
18
3,53%
Subjective questions
7
3,00%
8
1,57%
Free-choice items
0
0,00%
3
0,59%
Quantifier + NP
18
7,73%
4
0,78%
‘the others’
3
1,29%
0
0,00%
Species-generic use
8
3,43%
0
0,00%
Demonstrative strategies
1
0,43%
0
0,00%
‘one’
1
0,43%
0
0,00%
Person
1
0,43%
1
0,20%
23
Elliptical strategies
3
1,29%
0
0,00%
TOTAL
234
100.00%
510
100,00%
Table 5. ‘Other’ impersonal strategies in Dutch and Afrikaans
4.2.1 Specified nouns
Our Dutch and Afrikaans respondents frequently used nouns that denote a type
10
of individual
but do not refer to a particular person or particular people. Moreover, these types tie in closely
with the state of affairs expressed (e.g. ‘vandals’ with the act of stealing apples, ‘drivers’ with
the obligation to wear a seatbelt) and, therefore, they cannot really be said to identify anyone
in any more precise way. As (55) to (58) make clear, we find such nouns in universal as well
as existential uses in our data.
(55) Du: Vandalen hebben alle appels geplukt terwijl je sliep! (EXI-VAG-NN)
‘Vandals picked all apples while you were asleep.’
(56) Du: Alleen een formule 1 racer kan zo'n slipspoor maken, zeg! (EXI-INF-NN)
‘Only a formula one drives can have made such a skid mark, you know!’
(57) Afr: Kwajongens het sowaar al die appels gepluk terwyl jy geslaap het. (EXI-VAG-NN)
‘Rascals actually pick all the apples while you were asleep.’
(58) Afr: Alle motorbestuurders moet ten alle tye hulle sitplekgordel vasmaak.
(UNI-INT-NVER-MOD)
‘All drivers should fasten their seatbelts at all times.’
10
One of our reviewers suggested that this strategy should not be classified as an impersonal
strategy. While we understand their perspective, for the purpose of this study, we take a
broad approach to what we consider to be impersonal strategies. We include all strategies
that a speaker uses to avoid referring to a specific person or group of people. The lexical
meaning of these types of noun phrases makes it a possible strategy that speakers use to
refer categorically to a group of people without specifying it.
24
4.2.2 Imperative
Imperatives
11
may typically be used to issue directives to specific addressees. However, as (59)
to (61) show, they can serve to express obligations, prohibitions, permissions and recommen-
dations of a more generic type too, i.e. ones with which anyone who somehow feels that they
apply to them may comply (or not). Unsurprisingly, imperatives only occur in non-veridical
universal contexts in our data and, more specifically, mostly modal ones.
(59) Du: Poets je tanden, anders stink je de hele dag uit je mond. (UNI-INT-NVER-MOD)
‘Brush your teeth, otherwise your mouth will stink all day.’
(60) Afr: Maak nie saak wie jy is nie, maak altyd jou sitplekgordel vas! (UNI-INT-NVER-MOD)
‘No matter who you are, always fasten your seatbelt!’
(61) DU: Druk niet op die knop, want die is gevaarlijk! (UNI-INT-NVER-NMOD)
‘Don’t press that button, because it is dangerous!’
4.2.3 Relativisation
Our Afrikaans respondents sometimes combined a definite noun with the meaning ‘human be-
ing’, like die mense ‘the people’ in (62), or closely related semantics, like die ou ‘the guy’ in
(63), with a post-modifying relative clause that describes the state of affairs that the
11
One of the reviewers correctly noted that the presence of imperatives in our data may be an
artefact of our questionnaire design. As participants were presented with a bubble to fill in
with a possible utterance, they may have been more likely to choose an imperative expres-
sion than they would have in other methods, such as describing a picture. However, we
maintain that imperatives can still be considered impersonal strategies, as we instructed par-
ticipants at the beginning of the questionnaire to provide answers that do not refer to any
specific person or group of people.
25
individual(s) is(/are) assumed to be responsible. Within the entire sentence, this subordinate
clause can be argued to identify the referent(s) to some extent, but they are, in essence, still
unknown. We only find this pattern in existential uses, since it refers to a particular (group of)
individual(s) with certain characteristics.
(62) Afr: Die mense wat gister hier gedrink het, het hulle gemors net hier gelos. (EXI-INF-NN)
‘The people who were drinking here yesterday just left their garbage here.’
(63) Afr: Ek gaan die ou wat al die koek opgeëet het nóú 'n koekhou slaan! (EXI-SPE-PL)
‘I am going to hit the guy who ate all the cake.’
In such cases, our Dutch respondents consistently made use of free relative clauses instead, as
in (64).
(64) Du: Wie dit slipspoor maakte, reed veel te bruut!
‘[The person/people] who made this skid mark was driving way too aggressively!’
4.2.4 Infinitive
As non-finite verb forms, infinitives
12
do not require speakers to convey a first argument. They
can therefore be used, not unlike passives, to present a state of affairs as impersonal by leaving
it out altogether. As (65) and (66) show, the infinitive can serve this purpose in both universal
and existential contexts in both languages.
(65) Du: Het is nodig om meerdere soorten voedsel te eten (UNI-INT-VER)
‘It is necessary to eat different types of food.’
(66) Afr: Dink hoe gevaarlik dit is om só langs die pad te werk! (EXI-COR)
‘Just think how dangerous it is to work on the side of the road like that.’
12
One of the reviewers noted that infinitives is also very common in east Slavic languages
(Russian, Ukrainian) as an impersonal strategy.
26
4.2.5 Nominalisation
Our respondents used nominalised verbs, which do not need the first argument to be expressed
either, in the same way as infinitives. They too occur in universal uses, like het eten van rijst
‘the eating of rice’ in (67), as well as existential ones, like die gewerk aan die pad ‘the working
on the road’ in (68) in both languages.
(67) Du: Ik denk dat het eten van rijst in China op het dagelijks menu staat. (UNI-EXI)
‘I think that the eating of rice is part of the daily menu in China.’
(68) Afr: Die gewerk aan die pad mors my tyd! (EXI-COR)
‘The working on the road wastes my time.’
4.2.6 Agentive NPs
Another way that our respondents avoided an explicit impersonal first argument is by assigning
a certain agency to one of the other entities involved. By utilising theses type of strategy in
impersonal contexts, the respondents are able to refrain from explicitly identifying the agent of
the predicate. Consequently, according to our approach, these results can also be classified as
an impersonal strategy. In (69), for instance, the car may be the instrument but is portrayed as
the doer of the action and, hence, the actual (unknown) doer need not be expressed. Likewise,
the button in (70) is arguably the instrument (‘doing something with it’) or the patient (‘pressing
it’) but is presented here as causing something itself, removing the real doer from the picture.
A closely related strategy is the use of noun phrases that directly or indirectly imply the in-
volvement of one or more human beings. For example, in (71), there can only have been a
campfire if someone/some people lit it. In the same vein, the terror attacks in (72) cannot have
happened without actual terrorists. As is clear from (69) to (72), we find such cases again in
both universal and existential contexts.
(69) Du: Een auto heeft een flink slipspoor gemaakt. (EXI-INF-NN)
‘A car has made a serious skid mark.’
27
(70) Afr: Wat doen die knoppie? (UNI-INT-NVER-NMOD)
‘What does the button do?’
(71) Du: Nou, dat moet een flink vuurtje zijn geweest gisteravond (UNI-INF-PL)
‘Well, that must have been a nice little campfire yesterday.’
(72) Afr: Daar was terreuraanvalle op ñ paar plekke in die stad gewees (EXI-VAG-PL)
‘There had been terror attacks at a couple of places in the city.’
4.2.7 Subjective questions
Some of our visual stimuli depicted situations to which the respondents could formulate nega-
tive reactions (e.g. a reproach, an accusation, shock). For such cases, they occasionally used
what may be described as “subjective questions”: they ask which specific but unknown person
or people did something while evaluating them or the entire situation as negative, through the
use of negatively evaluative nouns and/or expletives. These questions were limited in our data
to existential uses such as (73) and (74).
(73) Du: Welke druiloor heeft onze computers gestolen? (EXI-VAG-NN)
‘Which numbskull has stolen our computers
(74) Afr: Wie de fok het al die koek opgeëet?! (EXI-SPE-PL)
‘Who the fuck has eaten all the cake?!’
4.2.8 Free-choice items
Our Afrikaans respondents sporadically employed free choice items for impersonal purposes.
Such items signal here that the interlocutors are at liberty to select who is intended: ‘no matter
who you choose from among …’ (see Vendler 1967: 80). With them, speakers can indicate that
they do not know or, in a sense, care which particular individual(s) is(/are) responsible, as in
existential (75), or that what they are saying applies to any person that you can think of, like in
universal (76). We did not come across any free choice items in our Dutch data but it is per-
fectly possible to produce utterances such as (75) and (76) in the language.
28
(75) Afr: Wie ookal gister hier gesit en drink het, het 'n groot gemors agtergelaat. (EXI-INF-NN)
‘Whoever was sat drinking here yesterday left a big mess.’
(76) Afr: Enige iemand wat ʼn motor wil bestuur, moet hul vasmaak met ʼn sitplekgordel.
(UNI-INT-NVER-MOD)
‘Anyone whatsoever who wants to drive a car should fasten their seatbelt.’
This set of strategies does not constitute a distinct impersonal strategy, as it closely aligns with
the strategy of employing indefinite pronouns in impersonal contexts (see section 5.1). How-
ever, we have chosen to still categorise these results under "other strategies" since it appears to
generate a less marked response when the FCI is incorporated, in contrast to its omission. As
such, we consider it a highly specific method of utilising indefinite pronouns alongside an FCI
in certain impersonal contexts. We acknowledge that additional research is necessary to gain a
better comprehension of the interplay between indefinite pronouns and FCIs as impersonal
strategies.
4.2.9 Quantifier combined with human noun phrase
Some respondents chose to make the parameter of quantification (see Section 2) explicit while
using a noun phrase with the meaning ‘human being(s)’. Such explication manifests itself in
our data through determiners like sommige ‘certain’ for existential uses and alle ‘all’ for uni-
versal uses. Consider the respective examples in (77) and (78). Note that we only came across
universal quantifiers for Afrikaans but that existential ones are an option too in the language.
(77) Du: waarom gooien sommige mensen hun afval naast de prullenbak! (EXI-INF-NN)
‘Why do certain people throw their garbage next to the trash can!’
(78) Afr: Alle mense moet te alle tye hulle veiligheidsgordel dra wanneer hulle bestuur.
(UNI-INT-NVER-MOD)
‘All people should wear their seatbelts at all times when they drive.’
This strategy is also closely related to another main impersonal strategy highlighted in section
5.1, namely NPHUMAN. However, the addition of a quantifier delineates the intended (albeit
unspecified) individual or group of individuals to which the respondent is referring. As this
29
represents an additional strategy that requires the respondent to specify the applicability of the
unspecified reference, we view it as a distinct technique that warrants separate discussion.
4.2.10 ‘Person’
The noun persoon ‘person’ denotes a human being
13
. Intuitively, it would therefore be a suita-
ble way to express an impersonal first argument – just like mens ‘human being’. Persoon is,
however, only found once in Dutch, i.e. (79) (see Section 4.2.13 about demonstrative deze in
the example), and once in Afrikaans, i.e. (80). The relative formality of the noun may play a
role here.
(79) Du: Zo, deze persoon weet wel hoe je een slipspoor maakt! (EXI-INF-NN)
‘So, this person does know how you make a skid mark!’
(80) Afr: Daar het sekerlik meer as 3 persone hier gedrink gister. (EXI-INF-NN)
‘More than 3 persons must have been drinking here yesterday.’
4.2.11 ‘The others’
A (very) small number of Dutch respondents employed de anderen ‘the others’ for specific-
existential contexts. As mentioned in Section 2, such uses are tied to the here and now of the
speech event and, as a result, the speaker may have certain ideas about who is responsible but
they are still not willing or able to identify this person or these people explicitly. De anderen
13
One of the reviewers questioned why we treat the use of persoon as a distinct strategy, rather
than categorising it under the 'NPHUMAN' strategy. We distinguish persoon as a separate strat-
egy because the 'NPHUMAN' category specifically encompasses nouns, whose lexical mean-
ing of 'HUMAN' has become grammaticalised to the point of being used as a pronoun. In
contrast, the nouns persoon or persone lexically refers to an unspecified individual, making
it a construction with inherent impersonal meaning.
30
in (81) suggests that the speaker indeed has suspicions about who ate the tart but they cannot
or will not say more than that a particular group of people not including themselves did it.
Although we did not find any cases of ‘the others’ in Afrikaans, it seems perfectly acceptable
in specific-existential contexts in this language too.
(81) Du: Ik had zo'n zin in taart en nu hebben de anderen de hele taart al opgegeten
(EXI-SPE-PL)
‘I was so looking forward to tart and now the others have already eaten the entire tart.’
4.2.12 Species generic use
For veridical universal-internal contexts, the Dutch respondents occasionally used nouns that
refer to the human species in general, like de mensheid ‘humanity’ in (82) and de mens ‘man’
(lit. ‘the human’) in (83). It is important to distinguish this type of noun from the main imper-
sonal strategy of NPHUMAN (see Section 4.1). The strategy we distinguish here does not concern
a grammaticalized impersonal pronoun or an indefinite set of people. Rather, the respondents
are attributing something to the whole of mankind. The two strategies are obviously related,
since nouns with the meaning ‘human being’ are known to start their grammaticalisation pro-
cess into impersonal pronouns in contexts where they have a species-generic meaning and refer
to all human beings (see Giacalone Ramat & Sansò 2007). Dutch men has this origin, coming
from man ‘man, human being’, and so does Afrikaans mens, which is in the process of devel-
oping into a full-fledged impersonal pronoun (see Van Olmen et al. 2019) but for which it is
therefore not always clear whether it has a species-generic or a truly impersonal interpretation.
We therefore just included it under NPHUMAN.
(82) Du: De mensheid heeft meer water nodig. (UNI-INT-VER)
‘Humanity needs more water.’
(83) Du: De mens heeft water nodig. (UNI-INT-VER)
‘Man needs water.’
31
4.2.13 Demonstrative strategies
An at first sight peculiar answer by one of our Dutch respondents, in (84), features demonstra-
tive die ‘those (ones)’. This item seems incompatible with impersonal contexts since its typical
function is to point to a particular known rather than unknown set of referents. It is not coinci-
dental, though, that this demonstrative is employed for an inferred-existential use. What ap-
pears to be happening here is that the speaker uses them to indicate the individuals whose
inevitable existence they have deduced from the available evidence. Die in (84) can be said to
point to the unknown group of people that the speakers assume must have been there to light
the fire of which the remnants are still visible. The argument that demonstratives actually fit
inferred-existential contexts quite well is supported by the fact that, for those uses, our Dutch
data also contains a small number of general nouns referring to humans (see, for instance, Sec-
tion 4.2.10) with a demonstrative determiner, as in (85). No demonstratives were found in our
Afrikaans data. The direct translations of (84) and (85) are, however, possible in this language
too for inferred-existential purposes.
(84) Du Die hebben gisteravond een lekker vuurtje gestookt. (EXI-INF-PL)
‘Those lit a nice little campfire yesterday evening.’
(85) Du Waarom gooien die mensen hun zooi naast de prullenbak? (EXI-INF-NN)
‘Why do those people throw their garbage next to the trash can?’
4.2.14 Elliptical strategies
Another (infrequent) set of answers restricted to our Dutch data is probably best described as
involving ellipsis. In (86), for instance, only the past participles gemaakt ‘made’ and gepar-
keerd ‘parked’ are present and no subject or auxiliary is included. The result, which may very
well be intended, is that it could be elliptical for a variety of other, often slightly more explicit
impersonal strategies: impersonal passive (er is) een slipspoor gemaakt (lit. ‘(there is) a skid
mark made’, number-neutral (iemand heeft) een slipspoor gemaakt ‘(someone has) made a skid
32
mark’, third person plural (ze hebben) een slipspoor gemaakt ‘(they have) made a skid mark’.
Interestingly, like demonstratives, such elliptical strategies were also only found for inferred-
existential uses. A very tentative hypothesis for this fact is that speakers use them to convey
the inferred state of affairs – such as having a barbecue and lighting a campfire in (86) – without
going as far as also explicitly indicating – through, say, mensen ‘people’ – the inferred exist-
ence of any individual(s) responsible for it. The reason, finally, why Afrikaans does not allow
the patterns in (86) and (87) in our view may have something to do with more general con-
straints in the language for ellipsis but is, at present, unclear to us.
(86) Du: zo'n slipspoor gemaakt en toch goed ingeparkeerd (EXI-INF-NN)
‘Made such a skid mark and still parked well.’
(87) Du: Barbecue gisteravond, vuurtje gestookt en niet gedoofd (EXI-INF-PL)
‘Barbecue yesterday evening, campfire lit and not extinguished.’
4.2.15 Een
We know from English (and other languages) that the numeral ‘one’ can grammaticalise into
a full-fledged impersonal pronoun. No such change has taken place in Dutch or in Afrikaans,
however. We nonetheless have one answer in our data, i.e. (88), where Dutch een ‘one’ seems
to occupy the subject slot of a conditional clause. Een may also be the indefinite article ‘a(n)’,
of course (numeral [en] and determiner [ən] are not distinguished in spelling), which we suspect
is the case here: the respondent must have forgotten to type the noun that was supposed to
follow ‘a(n)’. Yet, we do not want to exclude the possibility of impersonal een ‘one’ altogether.
It is a common phenomenon cross-linguistically and English influence, for instance, should
also not be written off completely as a potential factor. It is interesting to note in this regard
that, in Van Olmen & Breed’s (2018a) completion task, one of the Afrikaans respondents filled
in een ‘one’ too (the determiner is spelt differently, as ’n ‘a(n)’).
33
(88) Du: Als een een appel met wormen eet, groeit er dan een plant uit je oren?
(UNI-INT-NVER-NMOD)
‘If one eats an apple with worms, does a plant grow out of your ears?’
4.2.16 Summary
The overview in Section 4.2 shows that impersonal contexts can be and are, in fact, expressed
not only through established strategies like passives, personal pronouns, nouns meaning (or
deriving from) ‘human being’ and indefinite pronouns but also through a whole range of other
strategies. Most of them (e.g. infinitives, nominalisations) are found in both Dutch and Afri-
kaans. Moreover, even if they only occur in the data for one of these languages (e.g. ‘the oth-
ers’, free-choice items), it is nevertheless evident that they are an option in the other one too.
This overlap is obviously due the fact that Dutch and Afrikaans are closely related. There are,
however, also exceptions to this tendency, such as the use of elliptical strategies in inferred-
existential contents.
4.3 Discussion of the preferred strategies per context
Table 6 and Table 7 provide a summary, for Dutch and Afrikaans respectively, of all the strat-
egies used in the different impersonal contexts. The results are also presented in Figure 5 and
Figure 6.
We will limit ourselves here to a few general observations about the strategies that Dutch and
Afrikaans prefer for different impersonal contexts. A more in-depth discussion would require
more space than the present article allows and it would also only be really appropriate, espe-
cially for any comparisons between the two languages, if we took into account the Afrikaans
data for all visual stimuli (see Section 3.2).
A first observation concerns the use of personal pronouns. They are, by far, the dominant strat-
egy across all universal-internal contexts in Dutch and Afrikaans but figure less prominently
34
across the other impersonal contexts. This finding is in line with earlier ones for these two
languages (and others): Van Olmen & Breed’s (2018: 839) completion questionnaire, for in-
stance, also suggests that “the universal-internal domain has a much stronger preference for
pronominal forms of impersonalisation than the non-universal-internal one”. In addition, Afri-
kaans makes frequent use of imperatives in non-veridical-modal contexts and of (’n) mens par-
ticularly in veridical contexts, where it may still have its original species-generic meaning (see
Giacalone Ramat & Sansò 2007). For veridical uses, Dutch also appears to be partial to uni-
versal indefinite pronouns such as iedereen ‘everyone’.
35
Dutch impersonalisation strategies
PN.PRS
PN.INDF
men/mensen
Pasive
Spesified noun
imperative
Infinitive
Subjective Question phrases
Relativasation
Angentive NP
Nominalisation
Quantifier + NP
'others'
Species generic use
Demonstrative pronoun
'one'
'person(s)'
Elliptical
TOTAL
EXI-COR
36%
0%
2%
57%
0%
0%
0%
0%
0%
0%
5%
0%
0%
0%
0%
0%
0%
0%
100%
EXI-INF-NN
22%
21%
17%
8%
11%
0%
3%
5%
1%
7%
2%
1%
0%
0%
0%
0%
1%
2%
100%
EXI-INF-PL
26%
31%
13%
18%
4%
0%
1%
0%
2%
3%
0%
0%
0%
0%
1%
0%
0%
1%
100%
EXI-SPE-NN
13%
39%
0%
30%
10%
0%
0%
0%
0%
7%
1%
0%
0%
0%
0%
0%
0%
0%
100%
EXI-SPE-PL
51%
17%
1%
16%
10%
0%
0%
0%
1%
0%
0%
2%
3%
0%
0%
0%
0%
0%
100%
EXI-VAG-NN
8%
14%
1%
62%
13%
0%
0%
0%
0%
1%
0%
1%
0%
0%
0%
0%
0%
0%
100%
EXI-VAG-PL
5%
9%
0%
79%
5%
0%
0%
1%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-EXT
41%
18%
29%
7%
4%
0%
1%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-INT-NVER-MOD
63%
8%
4%
11%
1%
5%
6%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-INT-NEVER-NMOD
88%
1%
2%
1%
5%
1%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-INT-VER
53%
16%
15%
0%
1%
0%
2%
0%
0%
2%
0%
7%
0%
4%
0%
0%
0%
0%
100%
Table 6. Impersonal strategies in Dutch per context
36
Figure 5. Impersonal strategies in Dutch per context
37
Afrikaans impersonalisation strategies
PN.PRS
PN.INDF
men/mensen
Pasive
Spesified noun
Imperative
Infinitive
Subjective
Question phrases
Relativasation
Angentive NP
Nominalisation
Quantifier + NP
'others'
Species generic use
Demonstrative pronoun
'one'
'person(s)'
Elliptical
TOTAL
EXI-COR
55%
1%
3%
33%
0%
0%
4%
0%
1%
2%
1%
0%
0%
0%
0%
0%
0%
0%
100%
EXI-INF-NN
8%
15%
20%
24%
15%
0%
0%
0%
17%
0%
1%
0%
0%
0%
0%
0%
0%
0%
100%
EXI-INF-PL
19%
40%
1%
22%
11%
0%
0%
0%
5%
0%
0%
0%
0%
0%
0%
0%
0%
1%
100%
EXI-SPE-NN
39%
56%
3%
3%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
EXI-SPE-PL
34%
18%
1%
43%
1%
0%
0%
2%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
34%
EXI-VAG-NN
11%
48%
0%
18%
21%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
EXI-VAG-PL
6%
2%
0%
83%
7%
0%
0%
0%
0%
2%
0%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-EXT
23%
2%
30%
37%
5%
0%
1%
0%
0%
1%
1%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-INT-NVER-MOD
32%
7%
6%
11%
7%
36%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-INT-NEVER-
NMOD
79%
2%
1%
7%
12%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
UNI-INT-VER
34%
8%
50%
4%
0%
0%
0%
0%
0%
0%
3%
2%
0%
0%
0%
0%
0%
0%
100%
Table 7. Impersonal strategies in Afrikaans per context
38
Figure 6. Impersonal strategies in Afrikaans per context
39
Turning to the “semi-impersonal” contexts of universal-external and existential-corporate
(both contain clues that make identification not impossible), we can note that the first one ex-
hibits considerable variation in both Dutch and Afrikaans. What they share is the frequent use
of mense(n) ‘people’, which tends to be marginal for most other uses. This result suggests that
Haas’s (2018) findings for English people may apply to its Dutch and Afrikaans equivalents
too: it specialises in generic readings (e.g. ‘in China, people eat a lot of rice’) – compared to
the third person plural, which is more characteristic of episodic ones (e.g. ‘they have stolen all
the computers’). Ze and hulle ‘they’ still occur quite often for our universal-external stimuli,
however. A perhaps somewhat remarkable result (cf. [Anonymised] 2021: 196) for such con-
texts is the relatively high number of passives in Afrikaans. They are not found very frequently
in our Dutch data and, crucially, lack the explicit external perspective that ‘people’ and ‘they’
possess: ‘rice is eaten a lot in China’ can, in principle, include or exclude speaker and ad-
dressee. For existential-corporate uses, then, we can note that they have the same two dominant
strategies in both languages, i.e. the passive and the third person plural, even though Dutch
seems to favour the former and Afrikaans the latter. The passive arguably work well in such
contexts because it simply presents the state of affairs that itself already points to the entity
responsible for it. At the same time, ‘they’ fits very well too. The corporate character of the
referent is compatible with its original plurality as a proper personal pronoun and their semi-
identifiability with its original definiteness.
As regards the other existential uses, we can first of all observe that Dutch and Afrikaans are
both partial to the passive for vague ones. Yet, for number-neutral instances, the indefinite
pronoun iemand ‘someone’ is very common as well in Afrikaans. This finding suggests that, if
speakers assume that one unidentifiable person is responsible for something, they can (but need
not) signal this with their choice of strategy. The relative infrequency of ‘they’ in vague con-
texts in the two languages may be somewhat surprising, given its prominence in the literature
40
on impersonalisation. It might be taken to indicate that the third person plural’s original defi-
niteness (as well as plurality) is still felt to be present by a significant number of speakers of
Dutch and Afrikaans, who would then prefer not to use it to refer to an unidentified (group of)
individuals (see Van Olmen & Breed 2018). That said, ‘they’ does appear as a widespread
strategy for existential-specific contexts, in plural ones only for Dutch but in plural as well as
number-neutral ones in Afrikaans. An explanation for this phenomenon may be that respond-
ents are actually using the third person plural in a non-impersonal way here, to refer directly to
the people present in the here and now of the situation. Existential-specific uses also regularly
feature passives in Dutch and Afrikaans and, for number-neutral ones in particular, iemand
stands out in the two languages. Gast & van der Auwera (2013: 129) offer a reason for this
indefinite pronoun’s occurrence here: “[In existential-specific uses, t]here is a ‘physically pre-
sent’ and thus situationally accessible (singular or plural) agent, and a clearly perceptible event.
Situationally known/specific uses of impersonal pronouns are most similar to (quantifying)
indefinite pronouns like someone.” Existential-inferred contexts, finally, are generally the con-
texts in Dutch and Afrikaans with the least clear preference for particular strategies. We find
substantial numbers of passives, third person plurals, mense(n), indefinite pronouns, specified
nouns and relative strategies. Dutch especially exhibits a lot of variation there, with many “mi-
nor” strategies. Why the existential-inferred domain is so diverse is not immediately clear to
us but its comparative complexity (the referents are not simply unknown, their existence is
based on an inference and a state of affairs that itself is inferred) probably plays a role.
5 Conclusion
The main aim of this article was to determine what we can learn from using a visual question-
naire to investigate impersonal strategies in Dutch and Afrikaans (see Section 1).
41
In the first instance we learned or confirmed a number of things about the impersonal strategies
of Afrikaans and Dutch. We hope to have shown that, on the whole, the two languages have a
similar range of more established strategies (see Section 4.1) as well as less established ones
(see Section 4.2) at their disposal – although strategies unique to one of the two languages exist
too (e.g. elliptical strategies). We also hope to have shown that Dutch and Afrikaans share
certain preferences for specific strategies in particular impersonal contexts (e.g. favouring pro-
nominal forms of impersonalisation in universal-internal uses) but may also differ (e.g. univer-
sal-external passives in Afrikaans but not in Dutch) (see Section 4.3).
Furthermore, our study revealed that a visual questionnaire can be an effective method for
exploring language phenomena that are not frequently encountered in existing corpora. We
have discussed some weaknesses, including the not inconsiderable amount of unusable data –
due to the complexity of the functional domain under investigation – and the impossibility of
knowing whether personal pronouns are indeed intended impersonally – shared, at least to
some extent, with corpus studies (see Sections 3.2 and 4.1).
We nonetheless hope to have shown too that the open-ended, non-deductive character of the
method: (i) enables researchers to uncover a variety of strategies that languages use for imper-
sonalisation (e.g. not only more “expected” ones like nominalisations but also less “expected”
ones like assigning agency to a non-first argument; see Section 4.2) and would be hard, if not
impossible, to identify with other approaches; and, at the same time, (ii) confirms, from an
unbiased perspective, that the strategies studied most in the literature are also the most frequent
ones (see Section 4.1).
Of course, we do not wish to claim that corpus or questionnaire-based research is not necessary.
However, we believe that our deductive approach produces results that can then be the input
for other types of approach which rely on predetermined sets of impersonal strategies. An
42
acceptability judgments questionnaire based on data like ours could, for instance, subsequently
determine whether strategies are unacceptable in certain impersonal contexts, which is some-
thing that a corpus study or a visual questionnaire cannot do. In the same vein, a corpus study
could investigate how frequent the different strategies revealed by a study like ours are in actual
language usage.
To conclude, there are several ways in which the design and use of the visual questionnaire
could be further improved. One such way is to involve the respondents in a follow-up phase to
clarify any unclear or ambiguous answers. This can help to ensure that the data collected is as
accurate as possible. Additionally, as a questionnaire is designed based on the objectives of the
study, researchers have the opportunity to structure the questions and instructions in a way that
purposefully obtains the necessary answers. With these improvements, the use of visual ques-
tionnaires can provide valuable insights into language phenomena that may not be easily iden-
tified or investigated through other approaches.
6 References
[Anonymised] 2021a
[Anonymised] 2021b
Arif, Marina Mohd & Fatimah Hashim. 2008. Reading from the wordless: A case study on
the use of wordless picture books. English Language Teaching 1.121-126.
Bauer, Anastasia. 2021. Impersonalization in Slavic: A Corpus-Based Study of Impersonaliza-
tion Strategies in Six Slavic Languages. Journal of Slavic Linguistics 29. 123–178.
Coussé, Evie & van der Auwera, Johan . 2012. Human impersonal pronouns in Swedish and
Dutch: A contrastive study of man and men. Languages in Contrast 12.121-38.
43
Egerland, Verner. 2003. Impersonal pronouns in Scandinavian and Romance. Working Papers
in Scandinavian Syntax. Department of Scandinavian Languages, Lund University.
Hoekstra, Jarich. 2010. On the impersonal pronoun men in Modern West Frisian. The Journal
of Comparative Germanic Linguistics 13.31–59.
Garcia, Brigitte, Sallandre, Marie-Anne, L’Huillier, Marie-Thérèse & Aksen, Hatice. 2018.
Impersonal human reference in French Sign Language (LSF). Sign Language and Linguis-
tics 21.307–33.
Gast, Volker. 2015. On the use of translation corpora in contrastive linguistics: A case study
of impersonalization in English and German. Languages in Contrast 15(1). 4–33.
Gast, Volker & van der Auwera, Johan. 2013. Towards a distributional typology of human
impersonal pronouns, based on data from European languages. In Dik Bakker & Martin
Haspelmath (eds.) Languages Across Boundaries: Studies in the Memory of Anna Siew-
ierska, 119–58. Berlin: Mouton de Gruyter.
Groenen, Groenen. 2021. Hoe onpersoonlijk is persoonlijk?: Een contrastief corpusonderzoek
van de Afrikaanse en Nederlandse persoonlijke voornaamwoorden jy/je en hulle/ze.
Potchefstroom: North-West University. [MA Dissertation]
Haas, Florian. 2018. Changing human impersonal pronouns in English: A corpus study. Paper
presented at Societas Linguistica Europaea 51, August, Tallinn.
Horwat, Jeff. 2018. Too subtle for words: Doing wordless narrative research. Art/Research
International: A Transdisciplinary Journal 3.172–95.
Kirsten, Johanita. 2016. Grammatikale Verandering in Afrikaans van 1911-2010.
Vanderbijlpark: North-West University. [PhD]
44
Kitagawa, Chisato & Adrienne Lehrer. 1990. Impersonal uses of personal pronouns. Journal
of Pragmatics 14.739–59.
Luukka, Minna-Riitta & Raija Markkanen. 1997. Impersonalisation as a form of hedging.
Hedging and Discourse: Approaches to the Analysis of a Pragmatic Phenomenon in Aca-
demic, Raija Markkanen & Hartmut Schröder, Research in Text Theory, 168–87. Berlin:
De Gruyter.
Malchukov, Andrej L & Anna Siewierska. 2015. Impersonal Constructions: A cross-linguistic
perspective. Amsterdam: John Benjamins.
Marin-Arrese, Juana, Martínez-Caro, Elena & Pérez de Ayala Becerril, Soledad. 2001. A cor-
pus study of impersonalisation strategies in newspaper discourse in English and Spanish.
Paper presented at the Proceedings of the Corpus Linguistics 2001 Conference 29 March–2
April 2001, Lancaster University Centre for Computer Corpus Research on Language.
Mazzitelli, Lidia Federica. 2019. Referential and pragmatic-discourse properties of Lithuanian
reference impersonals: 2sg-imp, 3-imp and ma/ta-imp. Kalbotyra 72. 32-57.
Nodelman, Perry. 1990. Words about pictures. Georgia: University of Georgia Press.
Prenner, Maria Katarzyna, & Bunčić, Daniel. 2021. The Competition of ARB Constructions in
Polish. Journal of Slavic Linguistics 29(2). 201-220.
Primus, Beatrice. 2011. Animacy and telicity: Semantic constraints on impersonal passives.
Lingua 121. 80-99.
Sansò, Andrea & Anna Giacalone Ramat. 2007. The spread and decline of indefinite man-
strategies in European languages. An areal perspective. Paolo Ramat & Elisa Roma (eds.),
Europe and the Mediterranean as Linguistic Areas. Convergencies from a historical and
typological perspective, 95-131. Amsterdam: John Benjamins.
45
Siewierska, Anna & Papastathi. Maria. 2011. Towards a typology of third person plural imper-
sonals. Linguistics 49. 575-610.
Siewierska, Anna 2008. Ways of impersonalising: Pronominal vs verbal strategies. In María
de los Ángeles Gómez González , J. Lachlan Mackenzie & Elsa M. González Álvarez
(eds.), Current Trends in Contrastive Linguistics: Functional and cognitive perspectives, 3–
26. Amsterdam: John Benjamins.
Schlund, Katrin. 2018. A unifying approach to impersonality in Russian. Zeitschrift für
Slawistik. 63(1). 120-168.
Schlund, Katrin. 2020. Active transitive impersonals in Slavic and beyond: a parallel corpus
analysis. Russian Linguistics 44. 39-58. doi:10.1007/s11185-020-09221-2
Van Olmen, Daniël. & Breed, Adri. 2018a. Human impersonal pronouns in Afrikaans: A dou-
ble questionnaire-based study. Language Sciences 69.1-29.
Van Olmen, Daniël. & Breed, Adri. 2018b. Human impersonal pronouns in West Germanic: A
questionnaire-based comparative study of Afrikaans, Dutch and English. Studies in
Language 42.798 – 846.
Van Olmen, Daniël, Breed, Adri & Verhoeven, Ben. 2019. A corpus-based study of the human
impersonal pronoun (’n) mens in Afrikaans. Languages in Contrast 19.79–105.
Vendler, Zeno. 1967. Linguistics in Philosophy. Ithaca: Cornell University Press
Zobel, Sarah. 2015. A pragmatic analysis of german impersonally used first person singular
‘ICH’. In Barbara De Cock & Bettina Kluge (eds.), The referential ambiguity of personal
pronouns and its pragmatic consequences, Pragmatics, 26. 379-416.
