Content uploaded by Daniel Van Olmen
Author content
All content in this area was uploaded by Daniel Van Olmen on May 17, 2021
Content may be subject to copyright.
1
The passive as an impersonalisation strategy in Afrikaans and
Dutch: A corpus investigation
1 Introduction
Impersonalisation, i.e. “the process of filling an argument position of a predicate with a variable
ranging over sets of human participants without establishing a referential link to any entity
from the universe of discourse” (Gast & van der Auwera 2013: 136), has received considerable
attention in the recent literature (e.g. Malchukov & Siewierska 2011; De Cock & Kluge 2016).
An area that has attracted particular interest is the pronominal strategies that West Germanic
languages use for impersonalisation (e.g. De Hoop & Tarenskeen 2015; Auer & Stukenbrock
2018; Haas 2018a; Fenger 2018). Consider the examples in (1) and (2).
Afrikaans:
Mens leef net een keer.
one live only one time
‘One only lives once.’
Dutch:
Ze hebben mijn fiets gevonden.
they have my bike found
‘They have found my bike.’
In neither sentence does the subject refer to a specific, known individual or group of people. In
(1), the Afrikaans pronoun mens ‘one, you’ (deriving from the noun ‘human’) serves to make
a claim that, in principle, applies to everyone. In (2), the Dutch pronoun ze ‘they’ is used to
make a statement not about everyone but about the actions of a particular person or group of
humans who are still not explicitly identified.
Pronouns are, however, not the only strategies that languages have for impersonalisation (e.g.
Posio and Vilkina 2013; Gast 2015; Haas 2018b). They also need not be the preferred strategy
in certain impersonal contexts (see Van Olmen and Breed 2018a, 2018b and Section 2.2). Other
2
strategies include indefinite pronouns, generic noun phrases and agentless passives, as in (3) to
(5) respectively.
English:
Someone has stolen my bike!
Afrikaans:
Mense sê dit spook in daardie huis.
people say it haunt in that house
‘People say that house is haunted.’
Dutch:
De belastingen zijn weer verhoogd.
the taxes are again raised
‘Taxes have been raised again.’
Yet, all in all, little research has been undertaken on such non-pronominal impersonalisation
strategies. This comparative gap in the literature is what we aim to start filling with the present
article, by looking into the passive as an impersonalisation strategy in Afrikaans and Dutch.
1
On the basis of corpus data,
2
we would like to (i) determine if the passive is a productive
strategy for impersonalisation in both Afrikaans and Dutch, (ii) find out whether it is more
typically employed for some impersonal contexts than for others in the two languages and (iii)
compare its use as an impersonalisation strategy in Afrikaans to that in Dutch.
1
For clarity’s sake, we would like to stress the difference between the passive as an impersonalisation strategy
and the impersonal passive (see Kulikov 2011: 250-251). The latter term is typically reserved for constructions
that demote an active subject but do not promote anything to be the passive subject. Dutch er wordt gedanst ‘there
is dancing’ (lit. ‘there becomes danced’), with expletive subject er ‘there’, is a case in point. Such an impersonal
passive may be employed to avoid referring to everyone or to a specific, known person or group of people, i.e. as
an impersonalisation strategy, but need not function that way. In er wordt gedanst door de gasten ‘there is dancing
by the guests’, for instance, the human participants are clearly identified.
2
We refrain from calling this article a corpus linguistic study. Locating every single Afrikaans and Dutch passive
in existing corpora in an automatic way is difficult (if not unfeasible). Searches would need to be so broad that
finding the relevant hits amongst the “false positives” would be a never-ending task. Our queries (see Section 3)
therefore only target passives with a limited number of words between passive auxiliary and past participle. This
decision prevents us from making definitive quantitative claims about, say, the overall (relative) frequency (per x
number of words) of the agentless passive as an impersonalisation strategy in Afrikaans versus Dutch. Our data
will, however, still give us a good idea of, for instance, the different impersonal uses that the passive fulfils in the
two languages.
3
This article consists of four more sections. In Section 2, we will discuss the phenomenon of
impersonalisation in detail. Section 3 will introduce our methodological approach and corpus
data. In Section 4, the results will be presented and, in Section 5, we will give our conclusions.
2 Impersonalisation
3
It is already clear from (1) and (2) that impersonalisation is a multifaceted phenomenon. Mens
in (1) is roughly equivalent to ‘everyone’. This paraphrase is, however, not possible for ze in
(2), whose meaning is closer to ‘someone’ or ‘some people’. In fact, even more fine-grained
distinctions have been and need to be made. The people responsible for raising taxes in (5), for
instance, are probably not as “unidentifiable” as the person who stole the bike in (3). For an
analysis of the passive’s use as an impersonalisation strategy, a good understanding of the
different impersonal contexts is required. In what follows, we will therefore first describe the
seven criteria used by Siewierska & Papastathi (2011), Gast & van der Auwera (2013) and Van
Olmen & Breed (2018a, 2018b) to tell contexts apart and then introduce twelve possible
impersonal contexts.
2.1 Criteria
2.1.1 Quantification (UNI/EXI)
This criterion relates to whether the state of affairs applies to everyone relevant to or included
in the particular context – in which case it would be regarded as a universal (UNI) context – or
whether it is assumed that there is a specific but unidentified individual or group of people to
whom the state of affairs applies – in which case it would be considered an existential (EXI)
context. Examples (1) and (2) respectively illustrate this difference. Further examples are (6)
and (7).
3
This section is similar to the corresponding section in Breed and Van Olmen (subm.), from which most examples
here have been taken too. The two articles are part of a larger project about impersonalisation strategies in West
Germanic and the theoretical background is thus also essentially the same.
4
German: UNI
Man sollte nicht trinken und fahren.
one should not drink and drive
‘One should not drink and drive.’
English: EXI
They have increased the highway’s toll charge again.
2.1.2 Perspective (INT/EXT)
An impersonal context is said to involve an internal (INT) point of view when “a ‘center of
consciousness’ (e.g. the speaker or hearer) identifies, or is identified, with the set of referents
under discussion” (Gast and van der Auwera 2013: 139). When no such identification takes
place, the perspective is external (EXT). The difference can be exemplified by (8) and (9). The
speaker of (8) includes him- or herself in the context, imagining that he or she too would drink
a lot of Greek iced coffee when in Athens, and takes an internal point of view. By contrast, the
speaker of (9) does not include him-/herself in the context, indicating that it is the inhabitants
of Athens that drink a lot of frappé, and adopts an external perspective.
Dutch: INT
In Athene drink je veel frappé.
in Athens drink you much frappe
‘In Athens, you drink a lot of frappé.’
Dutch: EXT
In Athene drinken ze veel frappé.
in Athens drink they much frappe
‘In Athens, they drink a lot of frappé.’
2.1.3 Veridicality (VER/NVER)
The criterion of veridicality relates to the truth or realness of the state of affairs. When the
state of affairs is presumed to be true or is put forward as true, the context is veridical (VER).
Example (10) is a case in point: frequently being rained upon is presented as a fact for anyone
who visits or lives in England. When the state of affairs is only hypothetical, the context is
regarded as non-veridical (NVER), as in (11).
5
English: VER
In England, one often gets wet.
Dutch: NVER
Wat gebeurt er als je door een trilspin gebeten wordt?
what happens there when you by a daddy.longlegs bitten becomes
‘What happens when you get bitten by a daddy longlegs?’
2.1.4 Modality (MOD/NMOD)
Non-veridicality can be expressed through modal verbs. If this is the case for an impersonal
context, like in (12), where the state of affairs is not put forward as a fact but as something to
strive for, it is characterised as modal (MOD). An impersonal context is non-modal (NMOD)
when it is non-veridical but no modal form is used. Consider, for example, the conditional
subordinate clause as well as the interrogative main clause in (13).
Afrikaans: MOD
Jy moet altyd jou bes doen.
you must always your best do
‘You must always give your best.’
Dutch: NMOD
Leef je langer als je veel leest?
Live you longer if you much read
‘Do you live longer if you read a lot?’
2.1.5 (Un)knownness (COR/VAG/INF/SPE)
The amount of information regarding the (implied) unidentified human participant(s) available
to the speaker may differ in impersonalisation. More precisely, four types of “(un)knownness”
have been distinguished in the research: (i) corporate, (ii) vague, (iii) inferred and (iv) specific.
When it is possible to derive, from the state of affairs itself (and, more broadly, the discourse
context), who the particular person or group of people responsible for it is or are, even though
they are not explicitly named, the impersonal context can be characterised as partly known.
The typical situation in which this occurs is exemplified in (14). It involves a state of affairs
that can be assumed to be attributable to some kind of institution, which is why the context is
6
sometimes called corporate (COR). In (14), for instance, one can presume that it is the police
or some government agency that is setting up new speed cameras.
Dutch: COR
Ze zetten opnieuw flitspalen op.
they set again speed.cameras up
‘They are setting up speed cameras again.’
When it is not possible for the speaker to identify the particular person/people responsible for
a state of affairs, even though the speaker has specific information about it, the context is
regarded as vague. Sentence (15) can serve as an example: the speaker knows that the hearer’s
bike has been found and that the discovery is due to some specific (group of) individual(s),
about whom they are however unable to provide more detail.
Afrikaans: VAG
Hulle het jou fiets in die park gekry.
they have your bike in the park found
‘They have found your bike in the park.’
When the state of affairs is not actually known to the speaker but has been deduced – together
with the existence of the human participant(s) behind it – from some signs in the situation, the
context is regarded as inferred (INF). The speaker in (16), for instance, gathers from the smell
that an individual exists who made popcorn in the staff room.
English: INF
Someone has made popcorn in the staff room again, I can smell it.
Lastly, when the speaker is at the same time and place as the person/people accomplishing the
event there and then and may therefore have strong suspicions about who they are but does not
explicitly identify them, a context is specific (SPE), as in (17).
Dutch: SPE
Ah, ze kloppen op de deur.
ah they knock on the door
‘Ah, they are knocking on the door.’
7
2.1.6 Number (PL/NN)
The (implied) unidentified human participants in an (existential) impersonal context may be
necessarily plural (PL) or number-neutral (NN), i.e. either one person or more than one person.
This difference can be illustrated by (18) and (19): the action of gathering in the first example
always requires more than one individual whereas finding the hearer’s dog in the second one
could have been accomplished by one person or by a group of people.
Afrikaans: PL
Hulle het hier by mekaar gekom vir ’n partytjie
they have here by each.other come for a party
‘They have gathered here for a party.’
English: NN
They found your dog in the park.
2.1.7 Reportative evidentiality (SAV)
This criterion concerns the use of speech act verbs (SAV) such as ‘say’ and ‘claim’ to introduce
a statement as reported information coming from an unidentified and/or unidentifiable source
of evidence. Consider, for example, the use of the impersonal pronoun ‘they’ in (20) and its
not uncommon replacement ‘people’ in (21).
Afrikaans: SAV
Hulle sê eier is baie goed vir mens se hare.
they say egg is very good for one their hair
‘They say that egg is very good for one’s hair.’
Dutch: SAV
Mensen beweren dat hij schuldig is.
people claim that he guilty is
‘People claim that he is guilty.’
2.2 Contexts
Siewierska and Papastathi (2011) and Gast and van der Auwera (2013) employ the criteria
discussed in Section 2.1 in slightly different ways, to distinguish various impersonal contexts.
8
In Van Olmen and Breed (2018a, 2018b), the two approaches are integrated, resulting in twelve
contexts:
• four universal contexts (found in Gast and van der Auwera 2013);
• one corporate context (found in both Siewierska and Papastath 2011 and Gast and van
der Auwera 2013);
• one context with speech act verbs (found in Siewierska and Papasthati 2011);
• and six existential contexts, combining Siewierska and Papastathi’s (2011) criterion of
(un)knownness and Gast and van der Auwera’s (2013) criterion of number:
o vague contexts with a plural or a number-neutral reading;
o specific contexts with a plural or a number-neutral reading;
o inferred contexts with a plural or a number-neutral reading.
Table 1
4
provides a summary of these impersonal contexts and their discriminating features. It
is followed by an English example for each context.
Use
Quantification
Perspective
Veridicality
Modality
(Un)knownness
Number
Speech act verb
UNI-INT-NVER-NMOD
universal
internal
non-
veridical
non-
modal
N/R
N/R
N/R
UNI-INT-NVER-MOD
universal
internal
non-
veridical
modal
N/R
N/R
N/R
UNI-INT-VER
universal
internal
veridical
N/R
N/R
N/R
N/R
UNI-EXT
universal
external
N/R
N/R
N/R
N/R
N/R
EXI-COR
existential
N/R
N/R
N/R
N/R
N/R
N/R
EXI-VAG-PL
existential
N/R
N/R
N/R
vague
plural
N/R
4
N/R is short for ‘not relevant’ and is added whenever a criterion does not actually help distinguish one context
from another context.
9
EXI-VAG-NN
existential
N/R
N/R
N/R
vague
number-
neutral
N/R
EXI-INF-PL
existential
N/R
N/R
N/R
inferred
plural
N/R
EXI-INF-NN
existential
N/R
N/R
N/R
inferred
number-
neutral
N/R
EXI-SPE-PL
existential
N/R
N/R
N/R
specific
plural
N/R
EXI-SPE-NN
existential
N/R
N/R
N/R
specific
number-
neutral
N/R
SAV
existential
N/R
N/R
N/R
N/R
N/R
yes
Table 1. Twelve impersonal contexts
What happens if one breaks one’s little toe? (UNI-INT-NVER-NMOD)
One should not drink and drive. (UNI-INT-NVER-MOD)
You only live once. (UNI-INT-VER)
In Italy, they eat a lot of pasta. (UNI-EXT)
They have raised the electricity tariff again. (EXI-COR)
They have set off bombs in a few places in the city. (EXI-VAG-PL)
They have found your student card in the bathroom. (EXI-VAG-NN)
I see they had a party here last night. (EXI-INF-PL)
I can smell someone made popcorn in the staff room again. (EXI-INF-NN)
They are knocking on the front door and the back door at the same time. (EXI-SPE-PL)
They are waiting for you in front of your office. (EXI-SPE-NN)
People say that this place is haunted. (SAV)
3 Methodological approach
3.1 Background
Corpus research into impersonalisation strategies faces a number of challenges (see Van Olmen
and Breed 2018a: 6-7 for an in-depth discussion). A first one is that, with exceptions such as
the German pronoun man in (6), the strategies are typically not dedicated to the expression of
impersonalisation. Dutch je in (8), for instance, can serve as an impersonal pronoun as well as
a personal second person singular pronoun. In the same vein, passives may but need not serve
impersonalisation purposes (see fn. 3). Any corpus study of such strategies thus entails the
arduous task of separating the relevant from the irrelevant cases, often inevitably in a manual
way, and of dealing with potential ambiguity between impersonal and non-impersonal readings
(e.g. Siewierska and Papastathi 2011: 587-588 on the latter problem for ‘they’ in a variety of
languages; see also Section 3.2).
10
A second and related challenge is that, for many a strategy, the irrelevant hits for its specific
form will vastly outnumber the attestations that do fulfil the function of impersonalisation. It
seems fairly obvious, for example, that Afrikaans hulle occurs much more frequently as a
personal third person plural pronoun than as an impersonal pronoun, as in (15). There is, in
other words, a risk that, for certain impersonalisation strategies, possibly including the passive,
locating all the relevant instances in a sizeable corpus is a never-ending undertaking.
A third and final problem is specific to strategies like the passive, in Afrikaans and Dutch in
particular. Unlike impersonal pronouns, it involves not a single form that can easily be searched
for in an exhaustive manner but, as (34) shows for Dutch, the combination of an auxiliary –
with a range of forms – and a past participle – whose formation exhibits substantial variation
– with potentially considerable intervening linguistic material.
Dutch:
De olifant is/was/wordt/werd waarschijnlijk vanwege het erg slechte
the elephant is/was/becomes/became probably because.of the very bad
weer niet meer gefotografeerd/buitengelaten/benaderd.
weather not anymore photographed/let.out/approached.
‘The elephant hasn’t been/hadn’t been/isn’t/wasn’t photographed/let out/approached anymore
probably because of the very bad weather.’
Due to these features and especially for non-tagged corpora, any query whose results are more
or less guaranteed to include all passives (e.g. with any number of words between auxiliary and
past participle) is bound to be so open that many, if not most, of its hits will be false positives.
Sifting through those would simply not be feasible.
In view of the challenges mentioned above, this article will not present a comprehensive corpus
linguistic investigation of the passive as an impersonalisation strategy in Afrikaans and Dutch.
Rather, we will make use of samples of the results of more restricted queries (see Section 3.2).
However, the ease or difficulty with which these samples are compiled can still give us a good
idea of the productivity of the passive as an impersonalisation strategy. Moreover, the analysis
11
of the samples will enable us to determine in which impersonal contexts the passive occurs and
which ones are proportionally more or less frequent. Finally, our similar samples based on
similar Afrikaans and Dutch corpora will make a comparison of the two languages possible.
3.2 The data
To compile the comparable samples for our two languages, we drew on five corpora. For
Afrikaans, we used the 6-million-word fiction and 16-million-word newspaper parts of the
Taalkommissiekorpus (henceforth TK; Taalkommissie 2011) and two smaller corpora of
spoken language, i.e. the 560,000-word Gesproke Korpus van Afrikaans (Ponelis 1976) and
the 70,000-word Korpus van Gesproke Afrikaans (Van Rooy 2003) with mostly non-broadcast
interviews from the early 1970s and 2000s respectively (hereafter KGA for both). For Dutch,
we used the Lassy Groot corpus (LG; Van Noord et al. 2013) and, more precisely, a 6-million-
word subcorpus of fiction and a 27-million-word subcorpus of newspapers extracted from the
LG by Breed et al. (2017), and the Corpus Gesproken Nederlands’s (CGN; Nederlandse
Taalunie 2004) 560,000-word component with non-broadcast interviews form the late 1990s.
5
The sample for each language contains 600 passives serving as an impersonalisation strategy
and is made up in the following way:
• 100 simple tense passives from the spoken data (KGA and CGN)
• 100 simple tense passives from the newspaper data (TK and LG)
• 100 simple tense passives form the fiction data (TK the LG)
• 100 perfect tense passives from the spoken data (KGA and CGN)
5
We are aware that there is variation between the Afrikaans and Dutch corpora. The spoken data, for example, is
mostly from the 1970s for Afrikaans but from the 1990s for Dutch. The reason for this difference is the lack of
(access to) other more recent, sizeable speech corpora. Within such constraints, we have nevertheless chosen data
that is as similar in the two languages as possible (e.g. non-broadcast interviews).
12
• 100 perfect tense passives from the newspaper data (TK and LG)
• 100 perfect tense passives form the fiction data (TK and LG)
Two comments are in order here.
First, the variation in genres is motivated by the hypothesis that some impersonal contexts may
be more typical of one genre than of another. Speech and (the dialogues in) fiction will not
only give us a good idea of the passive as an impersonalisation strategy in “normal”
conversation but are also the places where one would expect, for instance, EXI-SPE cases to
occur, as they rely on the presence of unidentified human participants in the here and now of
the speaker’s situation (see Section 2.1.5). Such contexts seem much less likely to appear in,
say, newspapers. The latter genre, however, could be more concerned than regular conversation
with the dealings of institutions and hence be a more natural place for EXI-COR contexts to
arise (see Section 2.1.5). The different genres can thus be assumed to provide us with a more
comprehensive picture of the passive’s potential as an impersonalisation strategy.
Second, this aim of a fuller understanding also motivates the separate subsamples for simple
tenses (e.g. Afrikaans daar word gesê dat … ‘it is said that …’) and perfect tenses (e.g. Dutch
mijn telefoon is gestolen ‘my phone has been stolen’). The rationale behind them is that certain
impersonal contexts may be more characteristic of simple than of perfect tenses and vice versa.
It is not improbable, for instance, that UNI cases strongly correlate with simple tenses whereas
EXI-INF ones are more closely associated with perfect ones, as it seems more natural to make
inferences about the existence of a specific but unknown (set of) individuals based on signs of
their completed behaviour. Moreover, exploratory data indicates that perfect passives are much
less frequent than simple ones.
6
An entirely random sample of passives would therefore likely
6
Extrapolations of our samples suggest, for example, that, in spoken Dutch, simple passives have a frequency of
121.52 instances per 100,000 words while perfect ones occur just 35.03 times per 100,000 words. (These figures
were calculated as follows: dividing the number of all passives examined to reach the 100th impersonalising one
by the number of all randomized hits examined to reach the aforementioned point; applying this percentage to the
total number of query hits; dividing the resulting number by the subcorpus size; and multiplying it by 100,000).
13
have few perfect ones and, to some extent, obscure the passive’s potential in certain impersonal
contexts.
The data was extracted by searching for every past participle that features a form of one of the
passive auxiliaries (e.g. word and is in Afrikaans; word, werden, is and waren in Dutch) within
a span of one word to the right of it and four words to the left of it. The one-word rightward
range ensures that subclauses, for instance, are not simply excluded (e.g. Afrikaans … dat my
kar gesteel is ‘… that my car has been stolen’). The four-word leftward range allows for some
linguistic material to appear between auxiliary and past participle while avoiding too broad a
query that would produce an unworkable amount of false positives (see Section 3.1). The past
participles were looked for in two different ways. For the CGN, we relied on the part of speech
tag WW(vd,vrij,zonder), which is used for all past participles that do not serve as a modifier of
a noun or fulfil a nominal function (Van Eynde 2004: 31). For the other corpora, we employed
a formula that does as much justice to the variation in past participle formation as possible. The
Dutch LG corpus, for example, was searched for words beginning in ge-, be-, ont-, ver- or her-
(e.g. ge-zegd ‘said’, ver-deeld ‘distributed’) – potentially preceded by yet other morphemes
(e.g. dicht-ge-trokken ‘pulled shut’, uit-ge-voerd ‘carried out’) – and ending in -t, -d or -n (e.g.
ontdek-t ‘discovered’, gebakk-en ‘baked’).
After the extraction of the data, the results for each corpus were put in a random order and then
analysed manually until 100 passives serving as an impersonalisation strategy were found. This
process meant sifting out clearly irrelevant cases, including passives with an overt agent like
(35a) and entirely false positives like (35b). Bezit worden in (35b) matches our query but bezit
is a noun with the meaning ‘possession’ here and worden is not an auxiliary but the main verb.
Dutch:
a. … die deels door de overheid wordt gesubsidieerd.
… which partly by the government becomes subsidised
‘… which is partly subsidised by the government.’ (LG, fiction)
14
b. Daarom zal dit jullie erfelijk bezit worden.
therefore will this your hereditary possession become
‘Therefore, this will become your hereditary possession.’ (LG, fiction)
Moreover, not all agentless passives could be regarded as functioning as an impersonalisation
strategy either. When looking at (36) out of context, for instance, one might assume that this
sentence is a relevant hit.
Afrikaans:
Hy besef skynbaar wat van hom verwag word.
he realise apparently what of him expected become
‘He seems to realise what is expected of him.’ (TK, ficton)
However, when one considers what precedes in (37), it becomes evident that a particular known
human participant is implied, i.e. the woman bringing the bottles, and that the passive sentence
does not have an overt agent phrase as it can easily be derived from the context. In other words,
every agentless passive needed to be and was evaluated in view of its surrounding discourse.
Afrikaans:
Dan bring sy die bottels een vir een na die ou man se neus.
‘Then she brings the bottles to his nose one by one.’
… Hy besef skynbaar wat van hom verwag word.
… he realise apparently what of him expected become
‘… He seems to realise what is expected of him.’ (TK, fiction)
Another type of agentless passive that had to be excluded in light of its context can be illustrated
by (38). The sentences before gedwongen worden ‘be forced’ make clear that the first person
singular would be forced to write a letter not by a human participant but by their circumstances
of potentially running out of animal postcards. This passive does thus not meet our description
of impersonalisation in Section 1.
Dutch
Binnenkort zal mijn stapeltje dierenkaarten op zijn. Wat dan? Stuur ik je foto’s van
Provençaalse dieren? … Nee, als ik geen dierenfoto’s meer overheb,
‘Soon, my stack of animal postcards will have run out. What then? Will I send you photos of
Provençal animals? … No, if I do not have any animal photos left,
zal ik gedwongen worden je toch een echte brief te schrijven.
will I forced become you after.all a real letter to write
15
‘I will be forced to write you a real letter after all.’ (LG, fiction)
Unclear cases were resolved through discussion between the authors, of which the conclusion
occasionally was that an agentless passive was essentially ambiguous between an impersonal
and a non-impersonal interpretation. Consider, for example, the excerpt about an exterminator
in (39).
Dutch:
De meeste klussen kostten hem niet meer dan drie uur, van het zetten van de vallen tot en met
het sprayen. Hij spreidde zijn bezoeken over een paar weken zodat het meer werk leek dan het
was. Maar mensen waren bereid de hoogste bedragen neer te tellen om van ratten af te komen,
vooral als er ook kinderen in huis waren.
‘Most jobs took him no more than three hours, from setting the traps to spraying. He spread out
his visits over a couple of weeks so that it seemed like more work than it was. But people were
prepared to pay the highest prices to get rid of rats, especially if there were also kids in the
house.’
Wasberen, mollen en eekhoorns moesten diervriendelijk worden gevangen,
raccoons moles and squirrels had.to animal-friendly become caught
‘Raccoons, moles and squirrels had to be caught in an animal-friendly way,’
maar het kon bijna niemand iets schelen hoe ratten aan hun einde kwamen.
‘but almost nobody cared about how rats met their maker.’ (LG, fiction)
It is possible to read the passive here as an expectation specific to the exterminator of the story:
his clients want him to treat raccoons and the like in a humane manner. One can, however, also
interpret the sentence as a more general expectation applying to everyone, as the suitability of
replacing the passive with (40) and its impersonal pronouns suggests. For that reason, instances
like (39) were included in our sample.
Dutch:
Wasberen, mollen en eekhoorns moest je/men diervriendelijk vangen.
raccoons moles and squirrels had.to you/one animal-friendly catch
‘Raccoons, moles and squirrels, you/one had to catch those in an animal-friendly way.’
Note, finally, that the substitution test illustrated in (40) was employed more generally to check
whether an initial impersonal interpretation was likely or not in the context and, for the sake of
completeness, that we did not limit such replacements to impersonal pronouns but also at times
tested, say, an indefinite pronoun or a noun phrase like ‘people’.
16
3.3 Method of analysis
Each passive serving as an impersonalisation strategy was further analysed in terms of the
seven criteria discussed in Section 2.2, in order to determine the impersonal context in which
the passive is employed. To do this analysis, we first designed the decision tree in Figure 1, to
distinguish the twelve impersonal contexts.
Figure 1: Decision tree of impersonal contexts
We then used an IF statement with multiple AND/OR conditions
7
in Excel to analyse each
passive sentence according to the decision tree, as in Figure 2.
7
An example is the formula for the UNI-INT-NVER-NMOD context: =IF(AND(C2="universal"; D2="internal";
E2="nonveridical"; F2="nonmodal"; G2="NA"; H2="NA"; I2="NA"); 1; 0)
17
Figure 2: Screenshot of our analysis of passives in Excel
4 Results of the corpus investigation
In Section 1, we formulated three main research goals. Here, they are phrased as questions and
we will address them one by one in the following subsections.
4.1 Is the passive a productive impersonalisation strategy in Afrikaans and Dutch?
This question can be answered in various ways, depending on one’s definition of productivity.
We will address it here by examining whether or not the passive is used for the whole range of
impersonal contexts. Consider Table 2 and Figure 3 for the Afrikaans results, where the four
universal contexts are presented in purple and the eight existential ones in green, and Table 3
and Figure 4 for the Dutch findings, where blue indicates a universal context and yellow an
existential one.
18
Afrikaans
Simple passive
Perfect passive
Total
Newspaper
Spoken
Fiction
Newspaper
Spoken
Fiction
UNI-INT-NVER-NMOD
10
20
15
0
2
2
49
UNI-INT-NVER-MOD
17
17
9
1
0
0
44
UNI-INT-VER
0
1
1
0
1
1
4
UNI-EXT
8
3
1
0
4
1
17
EXI-COR
26
19
26
67
30
32
200
EXI-VAG-PL
4
9
5
5
4
8
35
EXI-VAG-NN
33
29
42
27
54
48
233
EXI-INF-PL
0
0
0
0
0
1
1
EXI-INF-NN
0
0
0
0
0
4
4
EXI-SPE-PL
0
0
0
0
0
0
0
EXI-SPE-NN
0
0
0
0
0
2
2
SAV
2
2
1
0
5
1
11
Total
100
100
100
100
100
100
600
Table 2: Afrikaans impersonalising passives and their uses
49 44
4
17
200
35
233
140211
0
50
100
150
200
250
Impersonal contexts
Afrikaans
UNI-INT-NVER-NMOD
UNI-INT-NVER-MOD
UNI-INT-VER
UNI-EXT
EXI-COR
EXI-VAG-PL
EXI-VAG-NN
EXI-INF-PL
EXI-INF-NN
EXI-SPE-PL
EXI-SPE-NN
SAV
Figure
3: Afrikaans impersonalising passives and their uses
19
Dutch
Simple passive
Perfect passive
Total
Newspaper
Spoken
Fiction
Newspaper
Spoken
Fiction
UNI-INT-NVER-NMOD
1
8
4
0
2
2
17
UNI-INT-NVER-MOD
18
15
26
0
0
0
59
UNI-INT-VER
7
5
4
0
0
0
16
UNI-EXT
6
4
8
0
0
0
18
EXI-COR
30
29
18
52
46
33
208
EXI-VAG-PL
9
12
10
16
13
6
66
EXI-VAG-NN
27
23
26
30
39
54
199
EXI-INF-PL
0
0
0
1
0
1
2
EXI-INF-NN
0
0
1
1
0
4
6
EXI-SPE-PL
0
0
0
0
0
0
0
EXI-SPE-NN
0
1
0
0
0
0
1
SAV
2
3
3
0
0
0
8
Total
100
100
100
100
100
100
600
Table 3: Dutch impersonalising passives and their uses
17
59
16 18
208
66
199
26018
0
50
100
150
200
250
Impersonal contexts
Dutch
UNI-INT-NVER-NMOD
UNI-INT-NVER-MOD
UNI-INT-VER
UNI-EXT
EXI-COR
EXI-VAG-PL
EXI-VAG-NN
EXI-INF-PL
EXI-INF-NN
EXI-SPE-PL
EXI-SPE-NN
SAV
Figure 4: Dutch impersonalising passives and their uses
20
Tables 2 and 3 and Figures 3 and 4 show that the passive is indeed used for almost the entire
range of impersonal uses. Consider the different universal types in (41) to (44) first. In (41),
the general claim that darning one’s winter socks – note the impersonal pronoun ’n mens! – is
fine is followed by a passive – in an (intrinsically non-veridical) conditional subclause – that
can be interpreted as applying to anyone, speaker and addressee included. This UNI-INT-
NVER-NMOD analysis can be confirmed by the appropriateness of the paraphrase ‘if one does
it too often’. The NVER-MOD context in (42) is UNI-INT too, as calling the woman’s hair
and hands beautiful is something that anyone is said to be able to do in any case.
Afrikaans: UNI-INT-NVER-NMOD
Daar is niks mee verkeerd om 'n mens se winterkouse te stop nie,
‘There is nothing wrong with darning one’s winter socks,’
maar as dit darem té veel gedoen word, lyk dit nie meer netjies nie.
but if it though too much done become look it not more neat not
‘but if it is done too often, though, it will no longer look neat.’ (TK, fiction)
Afrikaans: UNI-INT-NVER-MOD
Lank en skraal is sy, met 'n hoekigheid wat sedert haar huwelik al meer aan die versag is; selfs
haar bewegings het vloeiend geraak.
‘She is tall and slim, with a squareness that has been softening more and more since her
marriage; even her movements have become smooth.’
Twee dinge aan haar kan mooi genoem word sonder voorbehoud:
two things on her can beautiful called become without reservation
‘Two aspects of hers can be called beautiful without reservations:
haar dieprooi hare wat met die kleur van verouderde shiraz tot op haar skouers val, en haar
hande.
‘her crimson hair that drops to her shoulders with the colour of aged shiraz and her hands.’
(TK, fiction)
The newspaper article in (43), with a UNI-INT-VER passive, tells the story of a football player
whose mistake during a game made his manager consider taking serious action against him. To
describe what happened eventually, the journalist uses the common soup-centred saying below
expressing the universal fact that, in the end, one never takes measures as severe as those that
were initially announced. In (44), lastly, the writer is talking about the kissing habits in different
areas in France and states that people in and around Lille kiss four times. This generalization
21
does not appear to be intended as including speaker and/or addressee. When talking about the
custom in the South, the writer even explicitly uses third person plural ze – indicative of a UNI-
EXT perspective – rather than, say, UNI-INT je ‘you’.
Dutch: UNI-INT-VER
Een foute inworp en het daaropvolgende tegendoelpunt in de Champions League-wedstrijd op
Steaua Boekarest zou de directe aanleiding zijn geweest. Het klopt dat toenmalig Standard-
trainer Boskamp die avond in Roemenië razend was. …
‘A bad throw-in and the subsequent counter goal in the Champions League game at Steaua
Bucharest is said to have been the direct motivation. It is true that then Standard manager
Boskamp was furious that evening in Romania. …
Ach, de soep wordt nooit zo heet gegeten als ze wordt opgediend.
oh.well the soup becomes never so hot eaten as she becomes served
‘Oh well, the soup is never eaten as hot as it is served.’
Deflandre werd niet afgerekend op die ongelukkige laterale inworp.
‘Deflandre did not get written off on the basis of that unfortunate lateral throw-in. (LG,
newspaper)
Dutch: UNI-EXT
Aanvankelijk dacht ik dat het hier in het zuiden drie keer werd gedaan omdat de mensen meer
tijd hebben dan de altijd drukke Parijzenaars. Of misschien omdat ze meer mediterraan van
aard zijn, extraverter en emotioneler. Maar toen ontdekte ik dat
‘At first, I thought that, here in the south, it was done three times because people have more
time than the ever busy Parisians. Or perhaps because they are more Mediterranean in nature,
more extravert and more emotional. But then I discovered that’
in de omgeving van Lille, waar Alain geboren is, er vier keer wordt gezoend.
in the area of Lille where Alain born is, there four times becomes kissed
‘in and around Lille, where Alain was born, they kiss four times.’ (LG, fiction)
The various existential uses are exemplified in (45) to (51). The beginning of the article in (45)
features an EXI-COR passive. There exists a particular “entity” that is responsible for the state
of affairs but it is not overtly identified. We can gather from the verb itself, however, that this
entity is some kind of known institution with the power to fine people, like the police or the
judiciary. In (46), the claims are clearly made by a specific entity too and, to be precise, by a
group of people rather than an individual, given the plurality of the reports and the accusations.
Moreover, their existence is not inferred from the context and they are not present in the here
and now of the writer, making this a EXI-VAG-PL use. In (47), there are no inferences about
22
or immediate presence of any human participants either. The passive differs from that in (46),
though, in that the church may have been burgled by a group of people or by just one person.
It is, in other words, EXI-VAG-NN.
Afrikaans: EXI-COR
Riversdal. Die sakeman wat onlangs beboet is omdat hy in ’n natuurreservaat
Riversdale the businessman that recently fined is because he in a nature.reserve
gejag het,
hunted have
‘Riversdale. The businessman that has recently been fined because he was hunting in a nature
reserve’
het toestemming gehad om die wild in die reservaat te verwyder.
‘had permission to remove the wildlife from the reserve.’ (TK, newspaper)
Afrikaans: EXI-VAG-PL
'n Opposisieparty het gister die kop van dr. Monde Mayekiso, hoof van die departement van
omgewingsake se tak mariene en kusbestuur, geëis. … Dit volg op verskeie berigte die afgelope
ruk
‘An opposition party demanded yesterday that Dr Monde Mayekiso, head of the Department of
Environmental Affairs’ Maritime and Coastal Government branch, be sacked. … It follows
various recent reports’
waarin bewerings van swak en ontoereikende administrasie by mariene en
wherein claims of weak and insufficient management at maritime and
kusbestuur gemaak is
coastal.government made is
‘in which claims of weak and insufficient management at the Maritime and Coastal
Government have been made.’ (TK, newspaper)
Afrikaans: EXI-VAG-NN
Diefstal is 'n permanente gegewe in hierdie gewestes. Saam met vuur, mes, maan. Eer en liefde
ook.
‘Theft is a permanent given in these regions. Together with fire, knife, moon. Honour and love
too.
Daar is ingebreek by die Romaanse kerkie op die heuwel.
there is broken.into at the Roman little.church on the hill
‘They have broken into the little Roman church on the hill.’ (TK, fiction)
Next, both (48) and (49) can be regarded as EXI-INF: the occurrence of the state of affairs, as
well as the entity responsible for it, is a deduction in either case. In (48), the discarded stuff is
seen as evidence for the likely past event – note the epistemic necessity verb moeten ‘must’! –
of a group of people loading stolen bicycles and mopeds onto a truck. In (49), a woman’s
23
curious disappearance makes the first person narrator conjecture that someone may have
abducted her. The difference between the two examples is that the event in (48) necessarily
involves multiple human participants while that in (49) does not. It is interesting in this regard
that the sentence following the passive in (48) mentions multiple dieven ‘thieves’ whereas the
reply to the first person narrator’s guess in (49) brings up Daniel as a single suspect.
Dutch: EXI-INF-PL
De fietsen worden op bestelling gestolen. De bewoners van de appartementen aan de Rode
Kruislaan in Nijmegen hebben het spoor teruggevolgd. Buiten, onder de poort naar de
doorgaande weg,
‘The bikes are stolen to order. The residents of the apartments at Red Cross Lane in Nijmegem
tracked back the trail. Outside, under the gateway to the main road,
moeten de gestolen fietsen en brommers zijn ingeladen.
must the stolen bikes and mopeds are loaded.up
‘the stolen bikes and mopeds must have been loaded up.’
De voor de dieven overbodige spullen zijn daar namelijk teruggevonden .
‘The stuff that the thieves did not need was found there, you see.’ (LG, newspaper)
Dutch: EXI-INF-NN
‘Kunnen we misschien niet beter naar de politie gaan?’ probeer ik. ‘Hazelaar kon toch ook de
moord op uw man niet oplossen?’ antwoordt hij met een wedervraag.
‘‘Would it perhaps not be better if we went to the police?’ I try. ‘Hazelaar could not solve your
husband’s murder either, could he?’ he replies with a counter question.’
‘Maar misschien is ze wel ontvoerd!’
but maybe is she well kidnapped
‘‘But she may very well have been kidnapped.’’
‘Als het om de code gaat, dan moet Daniel erachter zitten.’
‘‘If it is about the code, Daniel must be behind it.’’ (LG, fiction)
Example (50) can be considered an instance of EXI-SPE-NN. Interviewee A is going on about
his students’ attitudes toward spelling mistakes when he is interrupted by a knock on the door.
He uses a passive to describe the action by some unknown individual at this specific time and
place before the recording is paused. Afterward, interviewer B attempts to get the conversation
going again. The SAV passive in (51), finally, features the speech act verb sê ‘say’ and presents
the following claim, which is supposed to substantiate the previous one, as reported information
but from some unidentified and/or unidentifiable source.
Dutch: EXI-SPE-NN
24
A: En dan zeggen ze ja maar we zien we zien op TV of we zien in de krant we zien wij daar ook
spellingfouten. Maar 'k zeg ik dan ... Ja. Si ze vinden dat zo precies niet meer belangrijk hè.
‘And then they say we see see on TV or we see in the newspaper we see spelling errors there
too. But I then say… Yes. They do not seem to find that so important anymore, right?’
Er wordt op de deur geklopt.
there becomes on the door knocked
‘There is a knock on the door.’
B: Ja. En dus ja. U u zegt u ziet daar uh ergens een evolutie in.
‘Yes. And so yes. You you are saying you can see some uhm evolution in there somewhere.’
(CGN, spoken)
Afrikaans: SAV
En omdat hy nie gevang word nie, miskien word hy aangemoedig of aangespoor om daarmee
voort te gaan.
‘And because he does not get caught, he is perhaps encouraged or incited to continue with it.’
Daar word gesê dat van die jongeres, kinders, word verslaaf op ʼn
there become said that of the young.people children become addicted in a
manier aan hulle eie adrenalien stimulus.
way to their own adreline stimulus
‘It is said that some of the young people, children, become addicted in a way to their own
adrenaline stimulus.’ (KGA, spoken)
The only context that is not attested at all for Afrikaans or Dutch is EXI-SPE-PL. A possible
reason for its absence is that EXI-SPE contexts in general, linked to the here and now of a
conversation, do not tend to arise in corpora. Siewierska & Papastathi (2011: 590) did not find
any EXI-SPE cases of ‘they’ in their nine-language translation corpus either: “[The identity of
the human participant(s) in this use] resides in the situational context …[, which] is manifest
to the speaker and not to the hearer, let alone the reader. One may therefore well imagine that
in written texts verbal contexts are much more likely to underlie 3pl IMP[ersonal] uses than
situational ones.” EXI-SPE can probably be assumed to be rare in the spoken language of
interviews too. There may be an immediate context but the genre and the formal character of
recording the session (for the corpus) makes it implausible that some action is undertaken at
the time and place of the interview by an unidentified individual, let alone a group of people.
To conclude the present section, we will briefly compare the passive’s impersonal uses to those
of the well/better-studied impersonal pronouns in Afrikaans and Dutch. As Table 2 and Figure
25
3 make clear, the passive is essentially employed in the whole range of impersonal contexts in
Afrikaans. In this respect, it is very different from the pronouns as impersonalisation strategies
in the language. Both (’n) mens and jy ‘you’ are restricted to universal uses with an internal
point of view while hulle ‘they’ is limited to UNI-EXT, SAV and existential contexts (see Van
Olmen & Breed 2018a: 19-20). The Dutch passive is similar to its Afrikaans counterpart when
it comes to its range of impersonal uses, as Table 3 and Figure 4 show, and so are the second
person singular je and the third person plural ze. Men, however, shares the passive’s potential
as an impersonalisation strategy of being able to occur in basically all universal and existential
contexts (see Van Olmen & Breed 2018b: 822, 838-839).
4.2 Is the passive more typically employed for some impersonal contexts than for others?
It is already clear from Tables 2 and 3 and Figures 3 and 4 that the answer to this question is
positive. In the present section, we will go into more detail. We will also investigate whether
the simple passive’s uses are similar to the perfect passive’s and whether any usage differences
exist between the subcorpora.
4.2.1 Universal versus existential contexts
Figures 5a and 5b demonstrate that impersonalising passives are more frequently employed in
existential contexts than in universal ones. In both Afrikaans and Dutch, existential uses make
up more or less 80% of our results.
26
This percentage is quite high compared to what the literature tends to say about the other well-
studied impersonalisation strategy that covers every use in many languages, i.e. the so-called
‘man’-pronoun (see Siewierska 2011; though, as mentioned before, its Afrikaans instantiation
(’n) mens is UNI-INT only). It is said to rarely serve existential purposes in actual usage (e.g.
Zifonun 2001 on German man ‘one, they’; Fonesca-Greber & Waugh 2003 on French on ‘one,
they’). For Dutch men and the passive, however, the contrast does not appear to be so stark, in
proportional terms. An exploratory examination of 75 arbitrary instances of men, where each
of our subcorpora makes up one third of the data, suggests that it is actually used existentially
in about 60% of the cases, with no substantial differences between the three genres. In absolute
terms, the existential passive is probably much more frequent than existential men, though, as
the impersonal pronoun seems to be disappearing from the language (e.g. Duinhoven 1990).
8
The numbers in Figures 5a and 5b could also be linked to earlier findings from a completion
task study (see Van Olmen & Breed 2018b: 819, 833-834). In this questionnaire, Afrikaans and
8
Support for this hypothesis comes from the comparison of men’s relative frequencies with extrapolations of our
sample data to the subcorpora in their entirety (i.e. determining the proportion of 100 impersonalising passives to
the number of randomized hits examined to reach the 100th relevant instance; applying this percentage to the total
number of query hits; dividing the resulting number by the size of the subcorpus; and multiplying it by 100,000).
Men and the impersonalising passive (would) occur 44 versus 260 times per 100,000 words in newspapers, 49
versus 96 times in spoken language and 38 versus 151 times in fiction.
114
486
0
100
200
300
400
500
600
Impersonal contexts
Afrikaans UNI vs EXI
UNI
EXI
Figure 5a: Afrikaans universal versus
existential impersonal contexts
110
490
0
100
200
300
400
500
600
Impersonal contexts
Dutch UNI vs EXI
UNI
EXI
Figure 5b: Dutch universal versus
existential impersonal contexts
27
Dutch native speakers were asked to fill in the subject slots of sentences representing, within
extra context, the entire range of impersonal uses. The syntactic slot to be completed clearly
favoured impersonal pronouns and, for the UNI-INT sentences, the participants indeed went
for impersonal pronouns in 95% of the cases in the two languages. Of the existential sentences,
just the EXI-COR ones exhibited a comparable proportion of impersonal pronouns in Afrikaans
(85%) as well as in Dutch (99%), which also had them in 84% of the EXI-VAG-PL ones. The
percentages for the other uses were far lower, indicating that “the universal-internal domain
has a much stronger preference for pronominal forms of impersonalization” (Van Olmen &
Breed 2018b: 839) than the rest of the impersonal domain. One way in which Figures 5a and
5b might now be interpreted is as tentatively suggesting that the passive (which the completion
task did not allow as an option) at least partially fills the existential gap left by the impersonal
pronouns in the two languages.
One crucial qualification regarding the preceding paragraph is that the ratio of universal versus
existential cases is highly dependent on the simple or perfect nature of the passive (see Section
3.2 too). Figures 6a and 6b provide, for Afrikaans and Dutch respectively, the simple and the
perfect passive’s proportions of universal versus existential uses.
28
The existential cases may outnumber the universal ones in both simple and perfect passives in
both Afrikaans and Dutch but, while the latter account for roughly one third of the simple ones,
they are hardly ever found in the perfect ones in either language. These results are, however,
not surprising. The perfect portrays an event as completed and this construal is hard to reconcile
with universal quantification. This type of quantification is concerned with, for instance, the
description of people’s typical behaviours, the presentation of general rules and expectations
and the expression of universal facts, which all tend to be unbounded. Examples (52) and (53)
are cases in point.
Afrikaans: UNI-INT-NVER-MOD
Ek is altyd baie bewus van die feit, jy weet, dat daar ʼn ou tradisie sou bestaan dat
‘I am ever very much aware of the fact, you know, that there would exist an old tradition that’
rooiwyn byvoorbeeld gedrink moet word teen kamertemperature.
red.wine for.instance drunk must become at room.temperatures
‘red wine, for instance, should be drunk at room temperature.’ (KGA, spoken)
Dutch: UNI-INT-VER
‘Ik heb nog steeds niet door hoe dat precies in z’n werk gaat, Muis,’ zeg ik timide. ‘Geeft niets,
luister, het werkt zo:’
‘‘I still do not understand how that works exactly, Mouse,’ I say timidly. ‘No problem, listen, it
works like this:’
met wit, dat wil zeggen officieel bestaand geld wordt iets gekocht.’
102
12
198
288
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
simple perfect
Afrikaans UNI vs EXI in
simple vs perfect
EXI
UNI
Figure 6a: Afrikaans universal versus
existential impersonal contexts in
simple versus perfect passives
106
6
194
294
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
simple perfect
Dutch UNI vs EXI in
simple vs perfect
EXI
UNI
Figure 6b: Dutch universal versus
existential impersonal contexts in
simple versus perfect passives
29
with white that wants say official existing money becomes something bought
‘with white or, in other words, officially existing money, one buys something.’ (LG, fiction)
It is nevertheless not impossible for a perfect passive to serve as a universal impersonalisation
strategy, as (54) with the rare impersonal pronoun man ‘man’ makes clear (if man was a noun,
the indefinite article would be required; see Van Olmen & Breed 2018a: 17 for details).
Afrikaans: UNI-INT-NVER-NMOD
As die roer op 200 ingestel is, hoeveel moet man vir 500 kompenseer?
if the rudder at 200 set is how.much must man for 500 compensate
‘If the rudder has been at 200, how much does one have to compensate for 500?’ (TK, fiction)
In this conditional context, the event in the subordinate clause – of setting the rudder of a gun
at 200 feet – must be finished before one can consider modifying the entire contraption in order
to compensate for a distance of 500 feet. The hypothetical completed nature of the first event
in relation to the second one makes a perfect passive acceptable here.
4.2.2 More specific impersonal contexts
Let us now go beyond the universal versus existential distinction and have a closer look at the
impersonalising passive’s more specific uses in Tables 2 and 3 and Figures 3 and 4. What we
have already commented on is the relative infrequency of the different universal-internal uses.
A preference for impersonal pronouns in these types of context has been offered as a potential
explanation. Other contexts that are (specially) rare are the EXI-INF and EXI-SPE ones: 1.16%
of the Afrikaans data and 1.50% of the Dutch data. The likely reason for their near absence has
also been discussed before, with respect to EXI-SPE-PL in particular (see Section 4.1). The
identity of the human participants in this use is anchored in the immediate situation evident to
the speaker but not the addressee, because of which it is probably unlikely to arise in written
language or in the structured and formal setting of recorded interviews. This argument can be
extended to EXI-SPE-NN, and to both EXI-INF uses. In the latter too, the inferred existence
of the human participant(s), as well as the inferred event itself, is very much situational. Note
30
in this regard that Siewierska & Papastathi’s (2011) nine-language translation corpus study did
not find any EXI-INF cases of ‘they’ either.
By far the two most common uses of the impersonalising passive in Afrikaans and Dutch are
EXI-COR – 33.33% of cases in Afrikaans and 34.67% in Dutch – and EXI-VAG-NN – 38.83%
in Afrikaans and 33.17% in Dutch. Their high frequencies are not a coincidence in our view,
as the passive seems particularly well-suited for them. First, what characterises EXI-COR is
the fact that the human participants, though not explicitly identified, are in a sense partly known
because the event described can only really be ascribed to a particular institution or entity. In
(55), for example, the act of sentencing someone can solely be done by the court.
Dutch: EXI-COR
Een jongeman van 22 uit Stekene is veroordeeld tot 240 uur gemeenschapsdienst.
a young.man of 22 from Stekene is sentenced to 240 hour community.service
‘A 22-year-old young man from Stekene has been sentenced to 240 hours of community service.’
(LG, newspaper)
Substituting the passive here for an impersonal pronoun such as ze would be perfectly possible
and acceptable and result in a clause like ‘they have sentenced a 22-year-old young man to 240
hours of community service’. A potential issue with the third person plural is that it is mostly
used as a personal pronoun and that addressees/readers may therefore be inclined to look for a
referent. The passive avoids this problem (which, admittedly, does not arise with men – though
this pronoun is disappearing and Afrikaans only has hulle as a pronominal option in EXI-COR
contexts) and, more generally, the inclusion of a marker whose reference is clear from the state
of affairs anyway.
Second, what is interesting about EXI-VAG-NN is that it occurs substantially more often than
EXI-VAG-PL in Afrikaans (six and a half times) and in Dutch (three times). A possible reason
for the passive’s partiality to EXI-VAG-NN is revealed by a more in-depth look at its corpus
examples. Strictly speaking, many of them cannot actually be considered “number-neutral”, in
31
that only one unidentified individual can be responsible for the event and never two people or
more. In (56), for instance, it is hard to see how the act of shooting a single guinea fowl with a
shotgun may have been accomplished by more than one person. In the vein of the “necessarily
plural” distinction made in the literature, one could analyse such cases as “necessarily singular”
or EXI-VAG-SG.
Afrikaans: EXI-VAG-NN
Dis ’n tarentaal wat een oggend kort na sonop in ’n bloekombos net
it.is a guinea.fowl that one morning short after sunset in a bluegum.forest just
buite Bergville op pad na Winterton geskiet is.
outside Bergville on road to Winterton shot is
‘It is a guinea fowl that was shot one morning just after sunset in the bluegum forest just
outside Berville on the road to Winterton.’
’n Haelgeweer is gebruik.
‘A shotgun was used.’ (TK, fiction)
Replacing the passive here by an impersonal pronoun is somewhat problematic. The only real
alternative in Afrikaans is the third person plural (it is probably also the most common one in
Dutch; see Van Olmen & Breed 2018b: 833). Hulle is known to be able to appear in number-
neutral cases (see Section 2.1.5 for an example). Its sense of plurality may therefore be argued
to have bleached to some extent in impersonal uses (see Van Olmen & Breed 2018a: 24). Yet,
in cases where the context is strongly suggestive of a single person, hulle remains strange, as
(57) shows (note that the ? judgment concerns the reading of the sentence as EXI-VAG).
Afrikaans: EXI-VAG-NN
? Dis ’n tarentaal wat hulle … geskiet het. ’n Haelgeweer is gebruik.
‘It is a guinea fowl that they shot … A shotgun was used.’
The passive is again a useful way of avoiding this problem, due to its lack of overt marking of
the plural/number-neutral/singular nature of the human participant(s). Another – undoubtedly
quite frequent – option for what we can call EXI-SG contexts would be the indefinite pronoun
iemand ‘someone’, of course (see also Van Olmen & Breed 2018b: 832-833).
It is important to add here that EXI-COR and EXI-VAG appear to be very typical of impersonal
32
‘they’ as well. In the earlier completion task study (see Van Olmen & Breed 2018b: 833-834),
EXI-COR was the context with the highest percentage of third person plurals in Afrikaans and
Dutch. EXI-VAG-PL and EXI-VAG-NN were among the contexts with the highest proportions
of ‘they’ too in both languages. The study did not look at EXI-VAG-SG but it is still interesting
to observe, in view of the preceding paragraph, that hulle and ze’s percentages decreased from
plural to number-neutral while the Afrikaans and Dutch passive is used more often in number-
neutral than in plural contexts. Siewierska & Papastathi’s (2011: 590) (limited) corpus data on
Dutch also show that EXI-COR and EXI-VAG are impersonal ze’s most frequent uses.
UNI-EXT and SAV, finally, are fairly marginal uses of the passive in Afrikaans (respectively
2.83% of cases and 1.83%) and Dutch (respectively 3.00% and 1.33%). The available evidence
for the third person plural, by contrast, indicates that they are quite normal for this impersonal
pronoun. Both are among the contexts with the highest proportions of ‘they’ in the Afrikaans
and Dutch completion task data and they are also attested multiple times, though half as often
as EXI-COR and EXI-VAG, in Siewierska & Papastathi’s (2011) corpus. Together, these facts
tentatively suggest that the two uses prefer pronominal to passive impersonalisation. We have
no immediate explanation why this would be the case for SAV. For UNI-EXT, however, one
reason may be that a passive like (58a) is essentially vague between an internal and an external
perspective. In the former interpretation, speaker and/or addressee can imagine themselves as
people visiting Italy and thus eating pasta. In the latter reading, the speaker is describing a habit
of the inhabitants of Italy.
Afrikaans
a. In Italië word pasta geëet.
in Italy become pasta eaten
‘In Italy, pasta is eaten.’
b. In Italië eet jy pasta.
‘In Italy, you eat pasta.’
c. In Italië eet hulle pasta.
‘In Italy, they eat pasta.’
33
This ambiguity may not always be desirable, though, and can easily be avoided by employing
impersonal pronouns rather than the passive: the second person singular or (’n) mens for UNI-
INT-VER and the third person plural for UNI-EXT, as in (58a) and (58b) respectively. In fact,
what our examples suggest is that passives are usually only used for UNI-EXT purposes when
the external perspective is evident from the context. Consider, for instance, the interview with
a Belgian radio presenter in a Belgian newspaper in (59).
Dutch: UNI-EXT
‘Wordt er stilaan niet te veel geleuterd op de radio?’
‘Are people not starting to waffle too much on the radio?’
‘In Nederland wordt er op de radio veel meer geleuterd dan hier.
In Netherlands becomes there on the radio much more waffled than here
‘In the Netherlands, people waffle much more on the radio than here.’
In alle landen rondom ons wordt vijf keer langer gepraat.’
‘In all countries around us, people talk five times longer.’ (LG, newspaper)
The entire setting itself already points to an external perspective. The explicit comparison of
the Netherlands to hier ‘here’, i.e. Belgium, (as well as the subsequent sentence) disambiguates
the passive even further as not internal but external: the interviewee is talking about a habit of
people on Dutch radio.
4.2.3 Comparison of the subcorpora
Table 4 and Figure 7 present the distribution of the various impersonal uses of the Afrikaans
passive in our three subcorpora: newspapers, spoken interviews and fiction. Table 5 and Figure
8 do just the same for the Dutch passive. In what follows, we will point out a few of the most
striking differences, mainly to show and stress that the passive’s use as an impersonalisation
strategy – like that of other forms of impersonalisation, undoubtedly – is not genre-independent.
A detailed discussion of all (dis)similarities is beyond the scope of this article, though.
34
Probably the most prominent contrast has to do with the passive’s two most frequent uses. In
Afrikaans, spoken language and fiction have approximately three EXI-VAG-NN instances for
Afrikaans
Newspaper
Spoken
Fiction
UNI-INT-NVER-NMOD
10
22
17
UNI-INT-NVER-MOD
18
17
9
UNI-INT-VER
0
2
2
UNI-EXT
8
7
2
EXI-COR
93
49
58
EXI-VAG-PL
9
13
13
EXI-VAG-NN
60
83
90
EXI-INF-PL
0
0
1
EXI-INF-NN
0
0
4
EXI-SPE-PL
0
0
0
EXI-SPE-NN
0
0
2
SAV
2
7
2
Total
200
200
200
Table 4: Uses of the impersonalising
passive in Afrikaans in newspapers
versus speech versus fiction
10
22
17
18
17
9
0
2
2
8
7
2
93
49
58
9
13
13
60
83
90
0
0
1
0
0
4
0
0
0
0
0
2
2
7
2
020 40 60 80 100
NEWSPAP E R S P O K E N FICTION
Afrik a a n s
SAV
EXI-SPE-NN
EXI-SPE-PL
EXI-INF-NN
EXI-INF-PL
EXI-VAG-NN
EXI-VAG-PL
EXI-COR
UNI-EXT
UNI-INT-VER
UNI-INT-NVER-MOD
UNI-INT-NVER-NMOD
Figure 6: Uses of the impersonalising
passive in Afrikaans in newspapers
versus speech versus fiction
Dutch
Newspaper
Spoken
Fiction
UNI-INT-NVER-NMOD
1
10
6
UNI-INT-NVER-MOD
18
15
26
UNI-INT-VER
7
5
4
UNI-EXT
6
4
8
EXI-COR
82
75
51
EXI-VAG-PL
25
25
16
EXI-VAG-NN
57
62
80
EXI-INF-PL
1
0
1
EXI-INF-NN
1
0
5
EXI-SPE-PL
0
0
0
EXI-SPE-NN
0
1
0
SAV
2
3
3
Total
200
200
200
Table 5: Uses of the impersonalising
passive in Dutch in newspapers versus
speech versus fiction
1
10
6
18
15
26
7
5
4
6
4
8
82
75
51
25
25
16
57
62
80
1
0
1
1
0
5
0
0
0
0
1
0
2
3
3
010 20 30 40 50 60 70 80 90
NEWSPAP E R S P O K E N FICTION
Dutch
SAV
EXI-SPE-NN
EXI-SPE-PL
EXI-INF-NN
EXI-INF-PL
EXI-VAG-NN
EXI-VAG-PL
EXI-COR
UNI-EXT
UNI-INT-VER
UNI-INT-NVER-MOD
UNI-INT-NVER-NMOD
Figure 5: Uses of the impersonalising
passive in Dutch in newspapers versus
speech versus fiction
35
every two EXI-COR ones (83 to 49 cases and 90 to 58 respectively). In newspapers, however,
this ratio is reversed (60 to 93 cases). Almost precisely the same difference is found in Dutch,
between newspapers on the one hand (57 to 82) and fiction on the other (80 to 51). The reason
for this higher frequency of EXI-COR in newspapers is probably that they simply write more
about issues involving institutions, which often do not need to be identified explicitly given the
event described, as in (60) and (61).
Afrikaans: EXI-COR
Hy is die naweek in die Moot-polisiekantoor aangehou voordat hy op borgtog
he is the weekend in the Moot-police.office detained before he on bail
vrygelaat is.
released is
‘He was detained at the Moot police station over the weekend before being released on bail.’
(TK, newspaper)
Dutch: EXI-COR
Masureel en Milosevic zijn geschorst.
Masureel and Milosevic are suspended
‘Masureel and Milosevic have been suspended.’ (LG, newspaper)
In (60), the institution involved in detaining him is of course the police while that responsible
for releasing him on bail is clearly the court. In (61), which appears in an article about football,
the two players mentioned here have obviously been suspended by the governing body of the
sport, the Royal Belgian Football Association.
Another, though small, difference worthy of note concerns EXI-INF and EXI-SPE. Such uses
may be very infrequent but it is, in our view, not a coincidence that all seven attestations in our
Afrikaans data occur in the fiction subcorpus and that seven of the nine cases in the Dutch data
are found in the fiction and speech subcorpora. As discussed in Sections 4.1 and 4.2.2, EXI-
INF and EXI-SPE are highly contingent on the immediate situation, making it unlikely that the
contexts arise in written language or in structured formal interviews. Still, unlike newspapers,
any spoken conversation between people will always retain the potential for these uses: there
is a here and now at which some unidentified individual can perform an action (EXI-SPE) or
36
from which a speaker can deduce the occurrence of an event and the existence of some person
responsible for it (EXI-INF). This argument can be extended to fiction, in that this genre may
contain imagined conversations with an immediate situation in the story, and explains why the
few EXI-INF and EXI-SPE cases in our data tend to occur where they do.
Other dissimilarities exist. Afrikaans newspapers, for instance, exhibit relatively more UNI-
INT-NVER-MOD cases than UNI-INT-NVER-NMOD ones (18 to 10), compared to spoken
language (17 to 22) and fiction (9 to 17). Dutch newspapers too have many more UNI-INT-
NVER-MOD cases than UNI-INT-VER-NMOS ones (18 to 1). We have no ready explanation
for this difference, however – like for other possible minor points of divergence between the
three subcorpora.
4.3 How does the Afrikaans impersonalising passive compare to the Dutch one?
It should be evident from the discussion in the preceding sections that the Afrikaans and Dutch
impersonalising passives resemble one another closely with regard to the parameters examined
here. The similarities between the two languages can be summed up as follows:
• The passive can be used in the whole range of impersonal contexts (only EXI-SPE-PL
was not attested in the data). In Afrikaans, it thus differs from the impersonal pronouns
as they are limited to either the universal-internal or the non-universal-internal domain.
In Dutch, the passive only shares its usage potential with men.
• The passive is employed much more frequently for existential than for universal purposes
– in its simple form and especially in its perfect form, whose aspectual construal clashes
with universal quantification, and compared to the only similar impersonal pronoun, men.
This fact may be due to a preference for pronominal forms of impersonalisation in the
universal domain.
• EXI-INF and EXI-SPE uses are very rare in our data and, in this respect, the passive is
37
similar to third person plural impersonal pronouns. The likely reason for the infrequency
of these contexts is that they are highly dependent on the speaker’s immediate situation,
which is not accessible in written language. They do tend to be slightly more typical of
the interviews and fiction than of the newspapers. This difference can be attributed to the
fact that, in the former subcorpora, there still exists a(n imagined) here and now allowing
actions by unidentified individuals or inferences about them.
• EXI-COR is one of the passive’s two most common uses. At least part of the explanation
must lie in its defining characteristic, i.e. of the event itself making the entity responsible
for it partially known. There may therefore be less of a need to indicate impersonalisation
with distinct markers such as pronouns (which, in the case of the third person plural, may
have the additional disadvantage of essentially being ambiguous between a personal and
an impersonal interpretation).
• EXI-VAG-NN is the passive’s other most frequent use and it occurs considerably more
often than EXI-VAG-PL. In this regard, the passive seems to differ from the third person
plural, for which the available evidence points to a preference for plural uses to number-
neutral ones. The high frequency of EXI-VAG-NN passives is probably in part motivated
by the fact that many EXI-VAG contexts are, in actual fact, necessarily singular. For such
cases, the alternative use of hulle or ze is somewhat incongruous: even in its semantically
bleached use as an impersonal pronoun, the third person plural clashes with the singular
nature of the event. This type of incompatibility does not arise with the passive (or with
an indefinite pronoun like iemand).
• The passive is regularly employed for UNI-EXT and SAV purposes. In view of what we
know about the third person plural’s uses, however, these contexts seem comparatively
uncommon. A potential reason for the relative infrequency of UNI-EXT is the passive’s
intrinsic vagueness between an internal and an external reading. Speakers might want to
38
avoid this ambiguity and use, respectively, jy, (’n) mens or je and hulle or ze to be explicit
about the perspective that they wish to communicate.
• Newspapers differ from the other subcorpora in that they contain more EXI-COR cases
than EXI-VAG-NN ones (but see below). The explanation for this difference is that the
former genre is bound to discuss the actions of institutions like the government, the police
and the judiciary more.
There are also dissimilarities between Afrikaans and Dutch. The ratio of EXI-VAG-NN cases
to EXI-VAG-PL ones, for example, is much larger in Afrikaans (six and a half times) than in
Dutch (only three times). The reason for this difference is unclear to us, though. Another point
of divergence between the languages has to do with the relation between the passive’s two most
frequent uses in spoken language. EXI-COR prevailing over EXI-VAG-NN has been said to
be characteristic of newspapers. Yet, it is also found – in a less extreme form – in our subcorpus
of Dutch speech (75 to 62 cases) but not in that of Afrikaans speech (49 to 83). The difference
is due to the nature of the two sets of interviews. The Dutch ones are all of school teachers and
centre around a more or less fixed series of questions about the impact of, among other things,
the 1995 spelling reform and education policies. As a result, many of the answers relate to the
actions taken by the institutions responsible for the domains being talked about (e.g. the Dutch
Language Union, the government). The Afrikaans interviews, by contrast, cover a wider range
of topics and interviewees and are more free-flowing.
5 Conclusion
In our view, Section 4.3, which is essentially a summary of this article’s findings, answers the
questions raised in Section 1, in showing that: (i) the passive is a productive impersonalisation
strategy in Afrikaans and in Dutch, being used in the entire range of impersonal contexts; (ii)
it is more typically employed for some impersonal purposes than for others – for instance, with
EXI-COR and EXI-VAG-NN as its top uses and with usage variation between different genres;
39
(iii) on the whole, the impersonalising passive behaves very similarly in the two languages.
The explanations offered for these results regularly refer to impersonal pronouns.
9
UNI-EXT’s
low frequency, for example, has been linked to the passive’s ambiguity in perspective and the
comparative explicitness of, say, mens versus hulle. In the same vein, to account for EXI-VAG-
NN’s high frequency, we have discussed the potential need to distinguish EXI-VAG-SG cases
and, subsequently, their apparent incompatibility with the third person plural. Our claims about
impersonal pronouns are, however, mostly based on data that is, strictly speaking, not wholly
comparable (e.g. very small corpus studies, completion tasks). The reason is that, thus far, no
systemic corpus investigations of the usage of Afrikaans and Dutch impersonal pronouns have
been undertaken. It is only such future research, though, that can really confirm, for instance,
whether explicitly external hulle is indeed employed more often than the vague passive for the
expression of UNI-EXT.
References
Auer, P. & A. Stukenbrock. 2018. When ‘you’ means ‘I’: The German 2nd Ps.Sg. pronoun du
between genericity and subjectivity. Open Linguistics 4, 280-309.
Breed, A., F. Brisard & B. Verhoeven. 2017. Periphrastic progressive constructions in Dutch
and Afrikaans: A contrastive analysis. Journal of Germanic Linguistics 29, 305-378.
Breed, A. & D. Van Olmen. subm. Developing and validating a visual questionnaire for the
study of impersonalisation strategies: A design thinking approach. Southern African
Linguistics and Applied Language Studies.
De Cock, B. & B. Kluge (eds.) 2016. Special issue: The referential ambiguity of personal
9
Another such explanation that should be added to the list is the use of the passive for existential purposes when
the context contains a potential referent for a third person plural pronoun. Speakers may wish to avoid impersonal
hulle or ze in this kind of situation – as the addressee could very well understand it as personal – and draw on the
passive (or men in Dutch) instead.
40
pronouns and its pragmatic consequences. Pragmatics 26(3).
De Hoop, H. & S. Tarenskeen. 2015. It’s all about you in Dutch. Journal of Pragmatics 88,
163-175.
Duinhoven, A.M. 1990. Verdwijnt men? In J.B. den Besten, A.M. Duinhoven & J.P.A.
Stroop (eds.), Vragende Wijs: Vragen over Tekst, Taal en Taalgeschiedenis. Amsterdam:
Rodopi, 70-80.
Fenger, P. 2018. How impersonal does one get? A study of man-pronouns in Germanic.
Journal of Comparative Germanic Linguistics 21, 291-325.
Fonesca-Greber, B. & L.R. Waugh. 2003. On the radical difference between the subject
personal pronouns in written and spoken European French. In P. Leistyna & C.F. Meyer
(eds.), Corpus Analysis: Language Structure and Language Use. Amsterdam: Rodopi,
225-240.
Gast, V. 2015. On the use of translation corpora in contrastive linguistics: A case study of
impersonalization in English and German. Languages in Contrast 15, 4-33.
Gast, V. & J. van der Auwera. 2013. Towards a distributional typology of human impersonal
pronouns, based on data from European languages. In D. Bakker & M. Haspelmath (eds.),
Languages Across Boundaries: Studies in the Memory of Anna Siewierska. Berlin: De
Gruyter, 119-158.
Haas, F. 2018a. “You can’t control a thing like that”: Genre and changes in modern English
human impersonal pronouns. In R.J. Whitt (ed.), Diachronic Corpora, Genre and
Language Change. Amsterdam: John Benjamins, 171-194.
Haas, F. 2018b. Changing human impersonal pronouns in English: A corpus study. Paper at
Societas Linguistica Europaea 51, Tallinn.
41
Kulikov, L. 2011. Passive to anticausative through impersonalization: The case of Vedic and
Indo-European. In A. Malchukov & A. Siewierska (eds.), Impersonal Constructions: A
Cross-Linguistic Perspective. Amsterdam: John Benjamins, 229-254
Malchukov, A. & A. Siewierska (eds.). 2011. Impersonal Constructions: A Cross-Linguistic
Perspective. Amsterdam: John Benjamins.
Nederlandse Taalunie. 2004. Corpus Gesproken Nederlands Release 1.0. The Hague: Dutch
Language Union.
Ponelis, F. 1976. Ponelis Gesproke Korpus van Afrikaans. Stellenbosch: Stellenbosch
University.
Posio, P. & M. Vilkuna. 2013. Referential dimensions of human impersonals in dialectal
European Portuguese and Finnish. Linguistics 51, 177-229.
Siewierska, A. 2011. Overlap and complementarity in reference impersonals: Man-
constructions vs. third person plural-impersonal in the languages of Europe. In A.
Malchukov & A. Siewierska (eds.), Impersonal Constructions: A Cross-Linguistic
Perspective. Amsterdam: John Benjamins, 57-89.
Siewierska, A. & M. Papastathi. 2011. Towards a typology of third personal plural
impersonals. Linguistics 49, 575-610.
Taalkommissie van die Suid-Afrikaanse Akademie vir Wetenskap en Kuns. 2011.
Taalkommissiekorpus 1.1. Potchefstroom: CTexT, North-West University.
Van Eynde, F. 2004. Part of Speech Tagging en Lemmatisering van het Corpus Gesproken
Nederlands. Leuven: University of Leuven.
Van Noord, G., G. Bouma, F. Van Eynde, D. de Kok, J. van der Linde, I. Schuurman, E.
Tjong Kim Sang & V. Vandeghinste. 2013. Large scale syntactic annotation of written
42
Dutch: Lassy. In P. Spyns & J. Odijk (eds.), Essential Speech and Language Technology
for Dutch: Theory and Applications of Natural Language Processing. Berlin: Springer,
147-164.
Van Olmen, D. & A. Breed. 2018a. Human impersonal pronouns in Afrikaans: A double
questionnaire-based study. Language Sciences 69, 1-29.
Van Olmen, D. & A. Breed. 2018b. Human impersonal pronouns in West Germanic: A
questionnaire-based comparative study of Afrikaans, Dutch and English. Studies in
Language 42, 798-846.
Van Rooy, B. 2003. Korpus van Gesproke Afrikaans uit die Vroeë 2000’s. Potchefstroom:
North-West University.
Zifonun, G. 2001. Man lebt nur einmal: Morphosyntax und Semantik des Pronomens man.
Deutsche Sprache 3, 232-253.
