In their article “Anaphor Binding: What French Inanimate Anaphors Show,” Charnavel and Sportiche (C&S) (2016) examine the distribution of two French anaphoric expressions: (a) the nonpossessive anaphor lui-même (feminine: elle-même; lit. ‘him/her/it-self’), and (b) the possessive anaphor son propre (lit. ‘his/her/its own’) when it induces focus alternatives on the possessor (henceforth, possessor son propre).1 On the one hand, C&S observe that, when these anaphors take an animate antecedent, their distribution is unaffected by the distance separating them from their antecedent: both anaphors can be linked to a local animate antecedent or to a more distant animate antecedent occurring in a different clause.2 On the other hand, C&S argue that, when these anaphors take an inanimate antecedent, they behave like “true anaphors” subject to locality restrictions that preclude any long-distance use. According to C&S, this generalization is supported by the existence of contrasts like those in (1) and (2).3
[Cette auberge]F.SG makes of shadow to its.3.M.SG
[Cette auberge]F.SG benefits of.the fact that the tourists
préfèrent soni (*propre) jardin à
prefer its.3.M.SG (own) garden.M.SG to
ceux des auberges voisines].
those.3.M.PL of.the inns neighboring
(adapted from C&S 2016:43, (12a,c))
[Cette loi]F.SG has caused the publication of-a book
[Cette loi]F.SG is so important that the journalists
prédisent la publication d’un livre sur ellei-(*même)
predict the publication of-a book on it.3.F.SG-(self)
et sur soni auteur.
and on its.3.M.SG author.M.SG
(adapted from C&S 2016:49, (25a), (26a))
In the (a) sentences, the anaphors son propre/lui-même and their inanimate antecedents belong to the same clause, and these sentences are natural. By contrast, in the (b) sentences, the same anaphors occur in an embedded clause while their inanimate antecedents occur in the matrix clause, and these sentences are degraded. Crucially, no such contrasts in acceptability are found when these anaphors take animate antecedents instead, as illustrated in (3) for son propre.
[Ce politicien]M.SG makes of shadow to his.3.M.SG
(propre) projet politique.
(own) project.M.SG political
[Ce politicien]M.SG benefits of.the fact that the élécteurs préfèrent soni (propre) projet
voters prefer his.3.M.SG (own) project.M.SG
politique à celui de soni adversaire].
political to that.3.M.SG of his.3.M.SG opponent.M.SG
C&S discuss further examples suggesting that, unlike their animate counterparts, inanimate son propre and lui-même are licensed only if their antecedents occur in their “local” domain, defined here as the smallest XP (e.g., TP, AP, DP) with an intervening subject containing these anaphors. On the basis of their observations, they propose that the distribution of these inanimate anaphors be captured by means of the following generalization (see C&S 2016:46–47):4
(4) Charnavel and Sportiche’s (2016) generalization Inanimate lui-même and possessor son propre must be syntactically bound by their (inanimate) antecedent within a local domain corresponding to the smallest XP with an intervening subject containing them.
[XP . . . Antecedent
*[XP . . . AntecedentXP Subject ≪ luii-même/sonipropre . . . ]]
C&S conclude that instances of French inanimate anaphors are always instances of plain anaphors and that, with some minimal amendment, Condition A of the binding theory (Chomsky 1986, 1993) offers an empirically adequate characterization of the conditions that regulate their anaphoric behavior. C&S further suggest that the locality restrictions imposed on these inanimate anaphors in turn support the idea that, in order to be exempt from Condition A, an anaphor must be linked to a “sentient” antecedent (i.e., one capable of thought, of having a point of view, of being an empathic target), in line with recent proposals according to which exempt anaphors are syntactially licensed by the presence of a logophoric or perspective center (Charnavel and Zlogar 2016, Charnavel 2017a,b).
In this squib, I report on novel data that challenge C&S’s conclusions by showing that inanimate son propre/lui-même are exempt from Condition A, but that their long-distance uses are subject to animacy- oriented blocking effects (Claim 1). Next, I show that animacy-oriented blocking effects also affect the distribution of animate son propre/lui-même (Claim 2). I will point out in conclusion how these blocking effects can be related to blocking effects of a similar kind previously described in the literature on long-distance anaphors.
1 Inanimate son propre/lui-même: Exemption from Condition A and Blocking Effects
The critical observation supporting Claim 1 is that long-distance uses of inanimate son propre/lui-même are in fact possible but sensitive to the properties of intervening subjects—specifically, to their animacy status. This new observation is illustrated by means of the contrasts in (5)–(8).
[Ce projet de recherche]M.SG of research has attracted some
sponsors]A qui veulent financer soni (*propre)
sponsors that want finance.INF its.3.M.SG (own) renouvellement.
‘[This research project]A that want to support itsi (*own) renewal.’
[Ce projet de recherche]M.SG of research has generated some
bénéfices]I qui peuvent financer soni (propre)
benefits that can finance.INF its.3.M.SG (own) renouvellement.
‘[This research project]I that can help support itsi (own) renewal.’
[La Terre]F.SG is mostly affected by the-activity
[des astronautes]A qui voyagent autour d’ellei- of.the astronauts that travel around of-it.3.F.SG- (*même).
‘[The Earth]A that travel around iti (*self).’
[La Terre]F.SG is mostly affected by the-effect of.the
satellites]I qui tournent autour d’ellei-(même).
satellites that revolve around of-it.3.F.SG-(self)
‘[The Earth]I that revolve around iti (self).’
[Chaque ordinateur]A each computer.M.SG requires some engineers
qui soient minutieux avec soni (*propre)
that are cautious with its.3.M.SG (own)
‘[Each computer]A that are careful with itsi (*own) operating system.’
[Chaque ordinateur]I each computer.M.SG requires some updates
qui soient compatibles avec soni (propre)
that are compatible with its.3.M.SG (own)
‘[Each computer]I that are compatible with itsi (own) operating system.’
[Chacun de ces deux termes]M.SG of these two terms is associated to a
sémanticien célèbre]A qui a contribué à sai
semanticist famous that has contributed to its.3.F.SG
(*propre) description linguistique.
(own) description.F.SG linguistic
‘[Each of these two terms]A that has contributed to itsi (*own) linguistic description.’
[Chacun de ces deux termes]M.SG of these two terms is associated to
[certaines restrictions sémantiques]I qui sont
certain restrictions semantic that are
directement liées à soni (propre) statut
directly related to its.3.M.SG (own) status.M.SG
‘[Each of these two terms]I that are directly related to itsi (own) linguistic status.’
In the (a) sentences, an animate subject intervenes between the anaphors son propre/lui-même and their remote inanimate antecedents, exactly as in examples (1)–(2), and these sentences are perceived as degraded, exactly as before. However, in the (b) sentences, these anaphors are now separated from their antecedents by an inanimate subject, and these sentences are accepted by speakers. For completeness, I note that these contrasts between animate and inanimate interveners disappear when the relevant intervener occupies a nonsubject position or when the target antecedent is instead animate. These additional observations are illustrated in (9) and (10), respectively (similar observations hold for lui-même).
[Ce cours d’eau]A vers this stream.M.SG of-water leads the swimmers toward
[un lieu]I qui est à l’opposé de sai (propre) a place that is at the-opposite of its.3.F.SG (own) source.
‘[This stream of water]A to [some place]I that is on the side opposite to itsi (own) source.’
[Ce cours d’eau]I vers this stream.M.SG of-water leads the waste toward
[un lieu]I qui est à l’opposé de sai (propre) a place that is at the-opposite of its.3.F.SG (own) source.
‘[This stream of water]I to [a place]I that is on the side opposite to itsi (own) source.’
[Ce politicien]A this politician.M.SG thinks that the voters
préfèrent soni (propre) projet politique à
prefer his.3.M.SG (own) project.M.SG political to celui de soni adversaire.
that.3.M.SG of his.3.M.SG opponent.M.SG
‘[This politician]A prefer hisi (own) political project to that of hisi opponent.’
[Ce politicien]M.SG thinks that the-economy of
marché]I favorise plus soni (propre) projet
market favors more his.3.M.SG (own) project.M.SG
politique que celui de soni adversaire.
political than that.3.M.SG of his.3.M.SG opponent.M.SG
‘[This politician]I favors hisi (own) political project over that of hisi opponent.’
Taken together, these data establish that inanimate son propre/lui-même can have long-distance uses, but that their long-distance anaphoric behavior is disrupted, and thus easily hidden, by certain intervention effects. Specifically, these anaphors can be separated from an inanimate antecedent by an inanimate subject, but the intervention of an animate subject triggers blocking effects. Given these new data, I propose that C&S’s generalization be revised accordingly.
Inanimate son propre/lui-même: Blocking effects
Instances of inanimate lui-même and possessor son propre are licensed only if no animate subject intervenes between them and their c-commanding antecedent. Intervening animate subjects trigger blocking effects.
[XP . . . Antecedent
*[XP . . . AntecedentXP SubjectA ≪ luii-même/sonipropre . . . ]]
[XP . . . AntecedentXP SubjectI ≪ luii-même/sonipropre . . . ]]
Having laid down this new generalization, let me now turn to some seemingly disproving data discussed by C&S. Specifically, C&S present a few examples of long-distance uses of inanimate son propre/lui-même like (12a–b) that do not involve the intervention of an animate subject and yet are perceived as unnatural or odd by native speakers, including those I consulted.5
[La Terre]F.SG suffers the fact that of numerous
satellites]I tournent autour d’ellei-(??même).
satellites revolve around of-it.3.F.SG-(self)
‘[The Earth]I revolve around iti (??self).’
[Cette montagne]F.SG attracts many of people
parce que soni (??propre) sommet est
by.this that its.3.M.SG (own) summit.M.SG is
l’un des sommets les plus escarpés du pays.
the-one of.the summits the most steep of.the country
(??own) summit is one of the steepest summits in the country.’
(C&S 2016:45, (18c); 43, (13b); judgments from my consultants)
Crucially, while the deviance of these sentences is expected on C&S’s generalization, it remains beyond the descriptive scope of (11). I argue, however, that this limit is desirable, for the deviance of these sentences follows from independent factors. Specifically, I argue that these sentences are perceived as odd because they fail to provide a suitable context for the felicity conditions associated with the contrastive flavor of son propre/lui-même to be met (see footnote 1). That is, the general contexts induced by these sentences do not allow speakers to entertain a plausible, discourse-coherent alternative that could contrast with the intended referent of son propre/lui-même, hence the resulting infelicity.
Three arguments support this line of explanation. First, these sentences are also perceived as odd when the relevant inanimate antecedents are replaced with animate ones, as in (13), thus showing that the animacy status of the antecedents is irrelevant here.
[Marie]F.SG suffers the fact that of numerous
enfants]A s’agitent autour d’ellei-(??même).
children wiggle around of-her.3.F.SG-(self)
‘[Marie]A are wiggling around heri (??self).’
[Cette musicienne]F.SG attracts many of spectators
parce que soni (??propre) piano est
by.this that her.3.M.SG (own) piano.M.SG is
l’un des pianos les plus beaux du pays.
the-one of.the pianos the most beautiful of.the country
Second, the inanimate anaphors in (12) become felicitous if we modify the surrounding context so as to provide a plausible, salient contrast, as in (14), thus showing that structural considerations are also irrelevant here.
[La Terre]I tournent autour d’ellei-(même).
‘[The Earth]I revolve around iti (self).’
Contrairement aux montagnes proches du Mt. Everest, [cette montagne]
‘Unlike the mountains near Mt. Everest, [this mountain]
Finally, these infelicity effects are also found with local instances of son propre/lui-même, as illustrated in (15a) for possessor son propre. Consistent with my explanation, these effects disappear if the context is adjusted appropriately—for instance, by making a discourse-coherent contrast explicit, as illustrated in (15b).
Jean.M.SG likes well his.3.M.SG father.M.SG but
he.3.M.SG prefers his.3.M.SG (own) brother.M.SG
Jean.M.SG likes well the brother of Marie but
he.3.M.SG prefers his.3.M.SG (own) brother.M.SG
Thus, I conclude that the effects in (12) are infelicity effects to be related to the (violation of the) contrastiveness condition imposed by the general focus properties of possessor son propre and lui-même, and are therefore orthogonal to the blocking effects unveiled in this squib and subsumed under (11).
2 Animate son propre/lui-même: Exemption from C-Command and Additional Blocking Effects
Thus far, we have seen that the distribution of animate son propre/lui-même is less restricted than that of their inanimate counterparts in that animate son propre/lui-même allow long-distance c-commanding antecedents regardless of the animacy status of the subjects intervening between them and their antecedents (recall (10)). However, there is another interesting difference in their distribution, which concerns their ability to take non-c-commanding antecedents. As C&S observe, inanimate son propre/lui-même requires a c-commanding antecedent.
(16) Inanimate son propre/lui-même: C-command requirement
[Ce problème]M.SG includes its.3.F.SG (own) solution.
[Les annexes de [ce problème]I incluent the appendices.M.PL of this problem.M.SG include sai (*propre) solution.
its.3.F.SG (own) solution.F.SG
‘[The appendices of [this problem]I include its (*own) solution.’
(adapted from C&S 2016:41, (6a–b))
On the other hand, animate son propre/lui-même allow non-c- commanding antecedents—for example, antecedents embedded in another NP.
(17) Animate son propre/lui-même: No c-command requirement
Marie.F.SG talks about her.3.F.SG (own) story.F.SG
[Les romans de [Marie]I parlent de sai
the novels.M.PL of Marie.F.SG talk about her.3.F.SG
‘[The novels by [Marie]I talk about heri (own) story.’
[Les couvertures des [romans de [Marie]I]I the covers.F.PL of.the novels.M.PL of Marie.F.SG
parlent de sai (propre) histoire.
talk about her.3.F.SG (own) story.F.SG
‘[The covers of [the novels by [Marie]I]I talk about heri (own) story.’
Now, the new observation that I would like to put forth in support of Claim 2 is that animate son propre/lui-même are sensitive to the animacy status of the NP(s) intervening between them and their non- c-commanding antecedents. Specifically, these anaphors can corefer with an embedded animate antecedent only if this antecedent is not contained in another animate NP. The paradigms in (18) and (19) establish this point. Both paradigms are built up similarly to the base paradigm in (17) by replacing the inanimate container NPs with animate ones.6
(18) Animate possessor son propre: Intervention of animate containers
Jean.M.SG follows his.3.M.PL (own) advice.M.PL
[Les voisins de [Jean]A suivent sesi
the neighbors.M.PL of Jean.M.SG follow his.3.M.PL
‘[[Jean]A follow hisi (*own) advice.’
[Les amis des [voisins de [JeanA]A
the friends.M.PL of.the neighbors.M.PL of Jean.M.SG suivent sesi (*propres) conseils.
follow his.3.M.PL (own) advice.M.PL
‘[[[Jean]A friends]A follow hisi (*own) advice.’
(19) Animate lui-même: Intervention of animate containers
Jean.M.SGNE-listens QUE him.3.M.SG-(self)
[Les voisins de [Jean]A n’écoutent que the neighbors.M.PL of Jean.M.SGNE-listen QUE luii-(*même).
‘[[Jean]A only listen to himi (*self).’
[Les amis des [voisins de [Jean]A]A the friends.M.PL of.the neighbors.M.PL of Jean.M.SG n’écoutent que luii-(*même).
NE-listen QUE him.3.M.SG-(self)
‘[[[Jean]A friends]A only listen to himi (*self).’
For completeness, I note that the long-distance use of these anaphors is also perceived as quite degraded whenever an animate subject intervenes between them and their embedded animate antecedent, as illustrated in (20) and (21).
(20) Animate possessor son propre: Intervention of animate subjects
[La déclaration de [Jean]I indique que the statement.F.SG of Jean.M.SG indicates that [cette bourse]I finance sesi (propres) this.F.SG grant.F.SG finances his.3.F.PL (own) recherches.
‘[[Jean]I indicates that [this grant]I supports hisi (own) research.’
[La déclaration de [Jean]I indique que the statement.F.SG of Jean.M.SG indicates that [cette politicienne]A finance sesi (??propres) this.F.SG politician.F.SG finances his.3.F.PL (own) recherches.
‘[[Jean]I indicates that [this politician]A supports hisi (??own) research.’
(21) Animate lui-même: Intervention of animate subjects
[La déclaration de [Jean]I indique que
the statement.F.SG of Jean.M.SG indicates that
[cette bourse]I ne finance que luii-(même).
this.F.SG grant.F.SGNE finances QUE him.3.M.SG-(self)
‘[[Jean]I indicates that [this grant]I only supports himi (self).’
[La déclaration de [Jean]I indique que
the statement.F.SG of Jean.M.SG indicates that
[cette politicienne]A ne finance que luii-
this.F.SG politician.F.SGNE finances QUE him.3.M.SG-(??même).
‘[[Jean]I indicates that [this politician]A only supports himi (??self).’
The core contrasts unveiled above can be descriptively captured as follows:
Animate son propre/lui-même: Blocking effects
Instances of animate lui-même and possessor son propre are licensed only if their antecedent NP is not contained within another animate NP.
*[NP . . . [NP Antecedent]A ≪ luii-même/sonipropre
[NP . . . [NP Antecedent]I ≪ luii-même/sonipropre
Together with the observations made above, the discovery of these blocking effects shows that the driving force behind the restrictions on inanimate son propre/lui-même also affects to some extent the distribution of their animate counterparts: both animate and inanimate son propre/lui-même are sensitive to the animacy status of the elements intervening between them and their antecedents, although their sensitivity manifests itself in distinct grammatical environments. Interestingly, we can now observe that in the end, the animacy-oriented blocking effects shown by son propre/lui-même closely resemble those previously found for other long-distance anaphors. To give just one example of this parallel, consider the paradigm in (23), based on examples (9)–(12) from Xue, Pollard, and Sag 1994 (see Tang 1989 and Huang and Tang 1991 for similar data), which exemplifies the animacy sensitivity of the anaphor ziji in Mandarin Chinese.
(23) Mandarin Chinese ziji: Animacy-oriented blocking effects
Zhangsan say Lisi harm.PERF self
[[Zhangsan]I haile zijii.
Zhangsan DE arrogance hurt.PERF self
‘[[Zhangsan]I hurt selfi.’
[[Zhangsan]I biaoming naben [shu]I
Zhangsan DE letter indicate this book haile zijii.
‘[[Zhangsan]I indicates that [this book]I hurt selfi.’
*[[Zhangsan]A haile zijii.
Zhangsan DE son hurt.PERF self
‘[[Zhangsan]A hurt selfi.’
*[[Zhangsan]I biaoming [Lisi]A haile
Zhangsan DE letter indicate Lisi hurt.PERF zijii.
‘[[Zhangsan]I indicates that [Lisi]A hurt selfi.’
The anaphor ziji can take local as well as long-distance c-commanding animate antecedents in subject position, (23a). In addition, ziji can take local as well as long-distance subcommanding animate antecedents, (23b) and (23c). Yet ziji also has animacy-oriented blocking effects: it can be anaphorically linked to a subcommanding antecedent only if no animate subject or animate NP container intervenes between it and its antecedent, (23d) and (23e). Overall, this pattern of blocking effects is reminiscent of the one we observed for inanimate and animate son propre/lui-même. Although the observed similarities need not call for a unified analysis, they invite us to envision the possibility that the sensitivity of these anaphors to the animacy status of intervening elements follows from a common linguistic or cognitive ground, and they give us a glimpse of how a more systematic comparison of the animacy-oriented blocking effects observed across languages could help us address this research question in future work.
3 Concluding Remarks
So what do French inanimate anaphors show? The bad news is: since these inanimate anaphors have local as well as nonlocal instances, their distribution provides no empirical evidence for a principled distinction between plain and exempt anaphors based on locality considerations and thus for the necessity of (some version of) the classical Condition A, contra C&S’s original proposal (see also Charnavel and Zlogar 2016). Yet the good news is: the distribution of these anaphors provides a new case study that can be used to further improve our understanding of the antecedence conditions on long-distance anaphors. For the time being, what French inanimate anaphors teach us is that animacy (i.e., sentience or aliveness) is a facilitating but not a necessary condition for exemption. On the one hand, this finding undermines the tentative generalization proposed in recent years that (the referent of) the antecedent of a long-distance anaphor must bear logophoric properties—for example, be capable of speech, thought, or consciousness (Charnavel and Sportiche 2016, Charnavel and Zlogar 2016, Ahn and Charnavel 2017, Charnavel 2017a, Charnavel and Huang 2018). On the other hand, it strengthens the view that the animate/inanimate distinction, with its possible cultural and subjective refinements, is to be treated on a par with other linguistic features and conceptual hierarchies (e.g., personhood, subjecthood), which have been found to shape the core grammar of long-distance anaphors in the world’s languages.
1 As discussed in Charnavel 2011, 2012 and C&S 2016, the anaphors lui-même and son propre tend to be emphasized in natural speech production and to have a contrastive flavor, especially in environments where they compete with their structurally simpler pronominal counterparts son/lui. In lui-même, the stress falls on même and induces alternatives on the associated pronoun, while son propre can yield different readings depending on its focus properties: in possessor son propre, the stress falls on propre, inducing focus alternatives on the possessor (e.g., ‘John used his OWN bike, not SUE’S’); by contrast, in possessum son propre, the stress falls on the head noun, inducing focus alternatives on the possessed (e.g., ‘John used his own BIKE, not his CAR’). This squib is concerned with the possessor readings of son propre since only those readings are argued in C&S 2016 to exhibit a correlation between binding locality and animacy status. All examples below are thus to be read in contexts that make the relevant alternatives salient.
2 It is worth noting that lui-même and son propre are not the only anaphoric expressions allowing long-distance uses in French. As Pica (1984a,b, 1986) shows, the third person reflexive soi (lit. ‘self ’) also has long-distance uses. The distribution of soi, however, is very different from that of lui-même and son propre: (a) the reflexive soi only takes animate antecedents, (b) it can only be used in generic statements, and (c) its long-distance uses are constrained by the Tensed-S Condition. In particular, Pica observes that, in a way similar to the Icelandic reflexive sig, French soi can be linked to a long-distance antecedent only if the embedded clause in which it occurs does not count as tensed (e.g., tenseless small clauses, sentences in subjunctive mood).
3 Throughout this squib, whenever relevant, I will indicate the animacy status of antecedents and intervening subjects using the superscripts I for inanimate and A for animate. Subscript indices are used to indicate the speaker’s belief state: two expressions α and β are coindexed just in case α and β are intended by the speaker to have the same denotation. Interlinear glosses use the following abbreviations: 3=3rd person, F/M=feminine/masculine, SG/PL=singular/plural.
4 Following C&S, I will consider that a subject α intervenes between inanimate lui-même/son propre and their c-commanding antecedent only if α is an intermediate c-commander. The use of “≪” in the schematic examples is thus intended to represent the c-command relation. See C&S 2016 for a discussion of this characterization in the case of DPs with a subject.
5 Specifically, C&S’s section 2, devoted to assessing the locality conditions on son propre/lui-même, includes 10 sets of examples that directly pertain to the contrasts discussed in this squib. Of these, 7 involve intervening animate subjects: (12), (14), (15), (16), (17), (19), (20) (sets of examples in other sections also involve intervening animate subjects: e.g., (25)–(26), (63b)). For space reasons, I restrict the present discussion to a couple of examples from the 3 remaining sets (i.e., (5), (13), and (18)); however, as far as I can tell, all these examples suffer from the same kind of infelicity discussed in the main text.
6 The relaxation of the c-command requirement is observed for other longdistance animate anaphors in a variety of languages: for example, Iron-Range English himself (Loss 2011), Mandarin Chinese ziji (e.g., Huang and Tang 1991, Xue, Pollard, and Sag 1994, Pollard and Xue 1998, Huang and Liu 2001, Cole, Hermon, and Huang 2006), Korean caki-casin (Kim 2000, Kim and Yoon 2009), Hindi/Urdu apnee (Davison 2001), Malayalam taan (Jayaseelan 1997, Swenson and Marty 2017), and Icelandic sig (Maling 1984). To the best of my knowledge, all these anaphors allow subcommanding antecedents (i.e., antecedents embedded within a (subject) DP that c-commands them) and exhibit restrictions very similar to those observed on animate son propre/lui-même: they can be anaphorically linked to a subcommanding animate antecedent as long as this antecedent is not contained in another animate NP. The case of Mandarin Chinese ziji is illustrated in the main text.
I thank Aurore González, Sophie Moracchini, Despina Oikonomou, David Pesetsky, and the LI reviewers, Squibs and Discussion editors, and copyeditor, for their thorough comments and suggestions, which significantly contributed to improving the quality of this squib.
Special thanks go to my language consultants for their time and patience. The judgments from French reported in the squib were collected through informal surveys with 8 native speakers of French (2 from Toulouse, 2 from Nantes, 2 from Paris, and 2 from Normandy). Those surveyed were presented with paradigms of sentences (e.g., minimal pairs), one at a time, displayed roughly as shown in this squib, and were asked to provide an acceptability yes/no judgment for each sentence. Overall, there was very little variation in acceptability judgments across the speakers surveyed. In terms of notation, I use “*” to indicate that those surveyed uniformly rejected a sentence as acceptable and “??” to indicate a nonperfect yet high rate of rejection (at most two yeses). The absence of those symbols indicates a null or quasi-null rate of rejection (at most one no).
This work was supported by the Bundesministerium für Bildung und Forschung (grant 01UG1411).