To p or to ¬pThe Bavarian Particle fei as Polarity Discourse Particle*
This article provides a description of the semantic and pragmatic conditions for the use of the
Bavarian German particle fei. I offer an analysis of fei as a discourse particle with strong
polarity focus1 effects. It is shown that the particle encodes polarity focus by contrasting a
proposition p with an alternative ¬p. Fei also acts like a discourse particle, since it works on
presuppositions in discourse, and does not affect the truth-value of a sentence.
Dieser Artikel beschreibt die Bairische Partikel fei, und die semantico-pragmatischen
Voraussetzungen ihrer Anwendung. Fei zeigt Polaritätsfokuseffekte, da sie eine Proposition p
mit ihrem Gegenteil ¬p kontrastiert. Fei muss aber auch als Diskurspartikel gelten, da sie
Präsuppositionen im laufenden Diskurs aufgreift und kommentiert, und auch den
Wahrheitsgehalt eines Satzes unverändert lässt.
1. The problem: what does fei mean?
Like Standard German (SG), the Bavarian-German dialect (BG) abounds with
discourse particles, also called ‘shading or toning particles’ (German: Abtönungspartikeln).
These particles, such as SG ‘auch, ja, doch, nur’ and BG ‘eh, aa, fei’, are notoriously
elusive in meaning, difficult to describe, and often lack a one-to-one correspondence in
translation. They are claimed to link an utterance to the larger discourse, i.e. previous
utterances, or even ‘comment’ on the unspoken. This could be a comment on the speaker’s
attitude, on the listener’s beliefs, or on the common ground (cf. Abraham 1991 and
references therein, Zimmermann 2007a/b). BG fei, analyzed in this paper, is such a particle.
Following I introduce the problem of capturing the meaning of fei. The ineffable
nature of discourse particles in general, and the problem of capturing the particular meaning
added by fei is illustrated below. As an answer to the question ‘What does fei mean?’,
speakers can answer with (1a) or (1b).
(1) a. Des is schwer zum
sog’n
that is difficult INF.DET2 say
‘That’s hard to say’
b. Des is fei schwer zum
sog’n
that is fei difficult INF.DET say
‘That’s hard to say’
When asked about the meaning difference between the two sentences, native BG speakers
relate that ‘it means about the same’. Yet, sentences with fei are generally judged to be
more emphatic than sentences without, and speakers report that the particle is used when
one wants to place emphasis on the thing said. (2) shows a similar contrast as (1). (2a) can
be uttered to someone who returns from the restroom with the fly unzipped; it does not
contain the particle. (2b) does, but the meaning, as in (1b), is altered in a subtle way.
*
I want to thank Hotze Rullmann, Martina Wiltschko, and Henry Davis for their invaluable comments and
their support along the way. Also thanks to the reviewers for helpful comments. Part of this research was
supported by a SSHRC (410-2005-0875) grant awarded to Hotze Rullmann.
1
For the purposes of this article I use the terms POLARITY and VERUM focus interchangeably.
2
The glosses are kept as simple as possible. The abbreviations used are INF=infinitive, DET=determiner,
PRT=particle PRON=pronoun, COMP=complementizer.
(2) a. Dei Hos’ntiarl is auf
your pant.door is open
‘Your fly is down’
b. Dei Hos’ntiarl is fei auf
your pant.door is fei open
‘Your fly is down’
Since the sentence pairs in (1-2) can be felicitously uttered in the given contexts, the
question about the exact semantic/pragmatic contribution of the particle arises. In the next
section I introduce my proposal for an analysis of fei. Evidence for this is presented in
section 3. Part 4 discusses some predictions made by the proposal. Finally, section 5
addresses how fei compares to the discourse particle doch, which has similar semantics.
2. Proposal: fei is a polarity focus discourse particle
In the remainder of this paper I argue that fei is part POLARITY FOCUS, part
it evokes a meaning along the lines of polarity focus (Höhle 1992),
and acts like a discourse particle by anchoring a host sentence p to a pragmatic context of
utterance, and by accessing the interlocutors’ epistemic attitudes toward p (Zimmermann
2007b, Fischer 2006). Like other discourse particles, and unlike some focus particles (cf.
Abraham 1991), the addition of fei does not affect the truth-functional meaning of a
sentence, but influences the felicity conditions in which a fei-modified proposition can be
uttered. In a nutshell fei expresses (3).
DISCOURSE PARTICLE;
This paraphrase of fei is the core of my proposal. I suggest that fei foregrounds the
opposite meaning of the proposition p it co-occurs with. It immediately explains the
emphasis that is connected with the use of the particle3: speakers use it when they want to
assert the polar opposite meaning of what the interlocutor seems to suggest or think. The
speaker reacts with fei to a proposition p, which either was uttered explicitly in previous
discourse, or was implied by the interlocutor’s non-verbal actions. The speaker corrects the
interlocutor with a “fei-focused” assertion: ”I am asserting p, since judging from your
behaviour/question/comment you seem to think ¬p, OR just in case that you think ¬p”.
3. Evidence
I show in (3.1) how fei adds polarity semantics to a sentence. A host proposition p is
contrasted with the alternative ¬p when the particle is used. (3.2) shows how polarity focus
alone cannot account for the data, and fei also needs to be considered as a discourse
particle. (3.3.) gives data evidence to support the analysis of fei as a polarity discourse
particle.
In that sense fei could be an overt marker of the speaker’s intent to update the common ground (Zimmermann
2007a).
3
3.1.
fei encodes polarity focus
I argue in this part that fei encodes polarity focus. This type of focus contrasts the
truth polarities of a sentence (Höhle 1992) by juxtaposing a proposition p with its polar
counterpart ¬p (or a proposition ¬p with the opposing p). The next paragraph explains the
concept of polarity focus, and spells out my assumptions about focus in general.
3.1.1. What is polarity focus?
Höhle (1992) proposes that polarity semantics is activated by VERUM, an abstract
element in the CP. It expresses a meaning along the lines of ‘It is true that p’, or it is the
case that p’. Given that VERUM is hosted in C, also the syntactic position of the finite verb
in German matrix clauses, an activation of VERUM (by stressing the verb) leads to the
observed polarity semantics. (4) illustrates.
Now consider the data in (5). Caps indicate stress.
(5)
A:
Hanna behauptet Karl schreibt einen
ROMAN.
Hanna states
Karl writes DET
novel
‘Hanna says that Karl is writing a NOVEL.
B:
Na, das stimmt.
Well that be.true
‘Well, that’s true.
Karl SCHREIBT einen Roman!
Karl writes
DET
novel
Karl IS writing a novel!’
In this dialogue two people discuss Karl’s writing a novel. B answers with a verum focused
matrix verb, realized by pitch accent. Höhle (1992) demonstrates this effect in German,
where VERUM is overtly realized as stress on whatever element occupies the head of CP
(see 4). In matrix clauses in SG and BG this position is occupied by the finite verb. (5)
shows that, somewhat unexpectedly, B can reply to A’s statement by focusing the finite
verb. B’s response is unexpected insofar as that a notion of focus, as proposed in Rooth
(1996), would predict that the lexical content of the verb schreiben ‘write’ is contrasted
with other elements of the same type, i.e. with other verbs. Focus in general is marked
phonologically, via pitch accent that marks a focused constituent (Jacobs 1992, Selkirk
1995). For Rooth, every expression [[X]] has an ordinary semantic value [[X]]0 and a focus
semantic value [[X]]F. The focus semantic value [[X]]F is a set of alternatives for the
focused constituent (marked by pitch accent, indicated by caps)(6).
(6)
[[Karl CAME]]0= [[Karl came]]
[[Karl CAME]]F= {Karl went, Karl wrote, Karl danced, Karl slept,…}
This example shows that the focus semantic value of the sentence ‘Karl CAME’ evokes the
set of all possible actions Karl did, i.e. ‘Karl X-ED’. Possible alternative focus semantic
values for [[came]]F is the set of activities Karl engages in: {went, wrote, danced,…}.
Thus, intuitively, a notion of focus is contrastive, in that it evokes alternatives. However,
for the specific example in (5), repeated as (7a), the pitch accent on the V2 verb does not
encode focus that evokes alternatives to the lexical content of the stressed verb.
(7) a. # B’ : ‘Karl x einen Roman’
b. # focus alternatives: {Karl liest ‘is reading’ einen Roman, Karl kauft‘ is buying’
einen Roman, Karl kritisiert ‘is criticising’ einen Roman, …}
Alternatives as in (7b) are grammatical, yet not felicitous in the dialogue given in
(5). The only felicitous possibility is polarity focus, an alternative type focus whose
contrast set contains only the proposition and its polar opposite (8)4.
(8) [[Karl schreibt einen Roman]]VF ={Karl schreibt einen Roman, ¬Karl schreibt einen
Roman}
Hence if it is asserted that ‘It is the case that Karl is writing a novel’, that proposition is
contrasted with ‘It is NOT true/the case that Karl is writing a novel’. Following, I show
that fei shows polarity focus effects as described in this section.
3.1.2. fei: polarity focus effects
I show here that fei displays focus effects, yet it is crucially different from other
focus particles, which associate with a particular element within a sentence and shift stress
(see also 3.1.3.). Following establishes that the focus projection of fei is necessarily the
whole sentence.
The Nuclear Stress Rule (NSR) for German (eg. Höhle 1982, Jacobs 1991), which I
adopt for Bavarian, has default sentential stress fall on the object, i.e. SEPP in example (9).
(9)
Da
Hans hod an
DET
Hans has DET
‘Hans (has) invited Sepp’
SEPP ei’glon
Sepp invited
Stress assigned with the NSR allows for the widest possible focus projection, i.e. up to CP.
In an alternative semantics framework, (9) could be theoretically contrasted by a seemingly
unlimited set of alternatives. Of course that set is in reality limited by discourse-pragmatic
Note that the contrast set of two members introduced by verum focus is in fact the same contrast set
introduced by a yes/no question. According to Hamblin (1973), questions are analyzed semantically as
introducing the set of possible answers. In the case of a y/n question this is a set of two propositions, based on
the question (see also Romero & Han 2001)
4
restrictions, but the focus projection itself allows for the full range of possibilities. The
addition of fei to (9) doesn’t shift the nuclear stress (indicated by caps), as observable in
(10), nor can fei itself be stressed, unlike e.g. ‘bloß’-’only’or ‘aa’-‘also’ (11).
(10)
Da
Hans hod fei an
DET
Hans has fei DET
‘Hans (has) invited Sepp’
(11)
Da
SEPP ei’glon
Sepp invited
Hans hod
FEI*/ BLOß/AA an
DET
Hans has
fei*/only/also DET
‘Hans (*fei)/ only/also invited Sepp’
Sepp ei’glon
Sepp invited
Nevertheless, even if no association with focus can be seen in fei sentences, a
particular focus effect with the addition of the particle is observable. Example (9) does not
exactly equal (10). Consider (12) which gives a possible context to the utterance (10):
(12)
Vroni and Christa are talking about an upcoming party.
Christa:
Da Hans
hod fei
an Sepp
DET Hans
has
fei
DET Sepp
‘Hans (has) invited Sepp’
ei’glon
invited
The focus alternatives to (12) must be a set which contains only the members in (13).
(13)
{Da Hans hod an Sepp ei’glon , ¬ Da Hans hod an Sepp ei’glon}
These limited focus alternatives for the fei sentence in (13) are clear in an extended context
(14), where another speaker (Maria) provides a retort. The focus alternatives (in the sense
of Rooth) given in (14 b-c) are infelicitous (marked by#). The former gives an alternative to
the object ‘Sepp’, the latter to the verb ‘invite’.
(14)
Vroni and Christa are talking about Hans’ party. Sepp is Hans’s best friend, so it
shouldn’t be a surprise that he is of course invited.
Vroni: Da
Hans hod fei
DET
Hans has fei
‘Hans has invited Sepp’
an
DET
Sepp ei’glon
Sepp invited
Maria: a: Ja
wos denn sunst?
Häd’an
NED eilõn
PRT
what PRT
otherwise
have.he.him not invite
‘Well then, what else? Should he NOT have invited him?
b: #Häd’a
an BÄDA
eilõn soin?
Have.he
DET Peter
invite should?
‘Should he have invited Peter?
suin?
should
c: #Häda
an
Sepp BUSSLN
Have.he
DET
Sepp kiss
‘Should he have kissed Sepp?’
soin?
should?
Maria’s answers in (14b-c) provide contrastive focus possibilities and are grammatical, yet
not felicitous in this context5. Only her offering the polar opposite (14a) as a retort is good
in this context. Also, note that unlike in the polarity focus cases discussed in (3.1.1), the
finite verb in fei sentences has no pitch accent associated with it, a requirement for the
“canonical” polarity focus cases. Nevertheless, a host proposition p is contrasted with some
alternative ¬p when fei is added, without a change in pitch pattern.
3.1.3. Fei: not a focus particle
This section illustrates that fei does not associate with focus (cf. Jackendoff 1972,
Rooth 1992), despite the polarity focus effects it displays, and thus it is not a focus particle
such as only (nur) or even (sogar). The following examples show that the addition of fei to
a host sentence with a stressed (i.e. contrastively focused) subject (15/17) or object (16/18)
does not shift the stress away from them. More importantly the focus-frame (bolded in the
examples) does not change between the sets (15/17) and (16/18), which it would with only
or even.
(15)Da BÄDA is a
Hund
(16)
DET Peter is DET dog
‘PETER is a great guy’= x is a great guy
Da Bäda is a HUND
DET Peter is DET dog
‘Peter is a GREAT GUY’=Peter is x.
An addition of fei has no effect on the contrastively F marked sentences of (15-16), i.e. it
does not shift the stress or change the interpretation.
(17) Da BÄDA is fei a Hund
(18)
DET Peter is fei DET dog
‘PETER is a great guy’=(fei)x is a great guy
Da Bäda is fei a HUND
DET Peter is fei DET dog
‘Peter is a GREAT GUY’= (fei)Peter is x.
In other words, the difference between (15) and (16) is the same as the difference between
(17) and (18), and the effects of free focus are unchanged with the addition of fei.
3.2.
fei as discourse particle
I claimed so far that fei encodes a meaning along the lines of polarity focus. In this
section I show that it encodes more than that. Consider (19), a conversation between two
people.
(19)
5
A:
I glab’ de Sechzga san recht beliebt in Minga.
‘I believe that 1860 (a soccer club) is pretty popular in Munich
These answers are felicitous if fei is not used.
B:
meist’n6 Muenchna san
fei Bayernfans
DET most
Munichers are fei fans.of. Bayern
‘Most Munichers are (actually) fans of Bayern München.’
De
In this example, B responds to A’s belief that 1860 is popular, with an utterance containing
fei. Under the polarity focus analysis, this example should be analyzed as introducing the
set of alternatives in (20).
(20) [[De meist’n Münchna san Bayernfans]]VF= {De meist’n Münchna san Bayernfans,
¬De meist’n Münchna san Bayernfans}
Nevertheless, the paraphrase that should follow from (20), i.e. “it is not the case that most
Munichers are fans of Bayern” is clearly not what B wants to express in her answer in (19).
The answer asserts that most inhabitants of Munich ARE in fact fans of the soccer club
Bayern München. Moreover, B’s answer implicates that A must not be aware of it. An
appropriate paraphrase of B’s answer is as follows in (21).
(21) I believe that you believe that it is not the case that most Münchner are fans of Bayern
Fei carries a presupposition, namely one about the addressee’s beliefs, thoughts, or
attitudes concerning the host utterance p. In this case it is the presupposition that A doesn’t
think that most inhabitants of Munich are fans of a particular soccer club. An answer
expressing only polarity focus is not what (19) expresses. B’s response to A would be
infelicitous if it accessed only the polar opposite to the proposition at hand. Thus an
explanation of the semantics of fei exclusively along the lines of polarity focus cannot be
the whole story, since it makes the wrong predictions about the data. There has to be
another meaning component in fei that renders the conversation in (19) felicitous, and still
captures the previous examples. In the following I show that the missing ingredient to a full
understanding of the particle, namely the speaker’s belief about the knowledge or beliefs of
the interlocutor, is to be found in the pragmatics of discourse particles.
3.3.
fei is a polarity discourse particle
Discourse particles express a speaker’s epistemic stance or attitude toward the
discourse context (Weydt 1969), in that they relate a host utterance to the shared common
ground (CG)(=what is presupposed) (cf. Kadmon 2000) between discourse participants.
One distinct feature I take to be valid for all discourse situations is that the CG does not
necessarily have to be verbalized (cf. Fischer 2006). I take “implicit propositions” that are
implicated by certain actions to be valid discourse antecedents suitable for updating the CG.
Also, unlike some focus particles, discourse particles do not affect the truth-functional
6
This example shows fei in the scope of a nominal quantifier. Fei here has (and always seems to have) scope
over the whole proposition. However, the question arises as to how conventional implicatures interact with
quantifiers. This is beyond the scope of this short paper and will have to await future research.
meaning of a sentence (Abraham 1991), but its felicity conditions. This also holds for fei
(see 3.1.3.).
I propose that fei points out a mismatch between the beliefs or knowledge of the
hearer and a proposition p the addressee seems to entertain; an utterance of fei α with
propositional content p is used in order to express the speaker’s assumption that the
addressee is actively entertaining ¬p and instructs the addressee to accept p as part of the
CG: ”I am asserting p, since judging from your behaviour/question/comment you seem to
think ¬p, or just in case that you think ¬p”.
(22) Context: A puts out an elaborate spread for dinner for 2 without uttering an invitation
p conventionally implicated by A’s setting the table for 2: Y OU HAVE TIME FOR DINNER
B:
I hob fei koa
Zeit mea
I have fei NEG.DET time more
‘I don’t have time any more’
Judging from your behavior I believe that you believe that I have time for dinner: but I don’t have time for
dinner.
In (22), the speaker reacts to her friend’s actions (setting the table for two) with a feifocused assertion, which can be paraphrased with “you don’t seem to realize, but I don’t
have time”. B’s use of fei points out to A a mismatch in what A assumes to be in the CG.
fei corrects that assumption with the polar opposite (having time vs. not having time).
(23) shows the same, this time with an explicit discourse antecedent in the form of a
question, that suggests that everybody is warm enough, so a window can be opened.
(23)
A:
Can I open the window?’
p implicated by A’s question: You are not cold
B:
Mi friats
Me is.cold
‘I am cold’
fei
fei
Judging from your question, I believe that you believe that I am not cold, but I am cold
Here fei ‘comments’ on the conventional implicature of the question (you are not cold),
correcting that implicature to the opposite.
There is, however, a use of fei that at first sight seems incompatible with my
proposal: fei can be used in some out-of-the-blue (OOB) cases, i.e. in contexts where no
explicit or implicit propositions are under discussion. The scenario in (24) illustrates:
(24) Context: First thing said by an aunt to her niece whom she hasn’t seen in a year.
“Du bist fei ganz schee g’waggs’n!“
you are fei whole nice grown
‘You have grown quite a bit’.
I assume that fei accesses more than explicit or implicit propositions entertained by the
addressee; fei accesses and can comment on readily observable facts as long as the speaker
thinks the interlocutor is unaware of that fact. Fei crucially cannot be used OOB if the
interlocutor is aware of what is asserted in the proposition (25).
(25) To a person who is shaking with cold:
#Di frierts fei.
you is.cold fei
To mean: ’I believe that you belive that you are not cold, but: you are cold’
Speaker comment: ‘Des ko ma ned sog’n, des merkt derjenige ja sejba’ (you can’t say that,
he notices that himself)
In the case of (24), the fact that she has grown is likely obvious to the addressee.
Nevertheless, the fei proposition can be uttered felicitously, since the addressee can
accommodate that the speaker believes she thinks ¬p. She was aware of p all along, but can
act as if she was not7. On the other hand, a shaking person cannot act as if they did not
realize they were cold, so (25) is unacceptable.
In this section I argued that the BG particle fei is best analyzed as a polarity
discourse particle. It shows polarity focus effects, and acts like discourse particle. An
utterance fei α with propositional content p is used in order to express the speaker’s
assumption that the addressee is actively entertaining ¬p and instructs the addressee to
accept p as part of the CG. It corrects an implicit or explicit proposition by picking out the
proposition with the opposite truth value.
4. Predictions of this proposal
The proposed analysis of fei as polarity discourse particle makes several predictions.
First, if fei accesses a presupposition and picks out the opposite polarity of that
presupposition, it should not be felicitous with y/n questions (4.1). Secondly, if VERUM is
encoded in C as proposed by Höhle (1992), and fei, like other discourse particles, is a
modifier on the propositional level (cf. Zimmermann 2004), fei should only be grammatical
in inflected imperatives with CP structure. Infinitival imperative forms with the particle
should be ungrammatical. This is borne out by the data (4.2).
4.1.
fei and Questions
The fact that someone asks a question implies that they do not know the answer to
that question (rhetorical questions aside8). A speaker asking a question does not update the
common ground with an actual proposition. Since fei corrects a proposition, the prediction
is that it should not be possible in an answer to a question. This is borne out both in y/n
questions (26) as well as in wh-interrogatives (27).
7
8
Thanks to a reviewer for pointing out that these cases can be handled in terms of accommodation.
Rhetorical questions often tend to be “masked” assertions. In these cases fei is a felicitous retort/answer.
(26) Question A:
Answer
Schneibts draussn?
‘Is it snowing outside?
B:
# Ja, es schneibt fei.
yes, it snows fei
# cannot mean ’I believe that you believe that it’s not snowing, but it actually is
(27) Question A:
Wer mog
an
Kafä?
‘Who wants coffee?’
Answer B:
#I mog fei an Kafä.
I want fei DET coffee
# cannot mean: ’I believe that you believe that I don’t want a coffee, but I actually do
Next, consider examples where fei is part of the question itself. In (28), the y/n
question with the particle is ungrammatical, as is the wh-question in (29). I take this to be
due to the same reasons as above; a question is not an assertion, and since fei corrects a
proposition (assertion), it is predicted to be only compatible with an utterance of the
appropriate illocution type.
(28)
#Schneibts
fei draussn?
Snows.it
fei outside
To mean ‘Is it snowing outside?’
(29)
# Wer mog fei an Kafä?
who wants fei DET coffee?
mean ‘Who wants coffee?’
Also consider (30), where the question A asks does carry a presupposition, yet that
presupposition is not the polar opposite of B’s answer; the resulting dialogue is infelicitous.
(30) Context: A asks ‘Who all here is not cold?’ (presupposes: someone is not cold)
B:
#Mi friats
fei neda
me be.cold fei NEG
‘I am not cold’
Compare this scenario with the felicitous (31).
(31) Context: A asks: Who all here is cold? (presupposes: someone is cold)
B:
MI friats
fei neda
me be.cold
fei NEG
‘I am not cold’
Note that the pronoun ‘MI’ needs to be focused (i.e. carry a pitch accent) for the answer to
be felicitous. If the pitch accent is missing, the discourse would be infelicitous. I assume
this to be due to the fact that with a contrastively focused pronoun, the question can be
construed as carrying the implicature that B in fact IS cold.
4.2.
fei in Imperatives
If polarity focus is encoded in CP, as suggested by Höhle (1992), and if discourse
particles are modifiers on the propositional level (Zimmermann 2004), the particle fei, as
encoding both, might be syntactically situated within the root CP. Bayer (1991) states that
discourse particles encode attitudinal meaning that cannot be activated unless the particle
has access to a root clause (i.e. CP). It is thus predicted that fei cannot occur in structures
that lack the CP layer, such as infinitival commands. SG and BG have two ways of forming
commands: ‘true’ imperatives, and infinitival commands. Only the imperative, as shown in
(32), has the verb move to C. Evidence for the presence of a CP layer is the separableprefix verb ‘auf-bassn’, which is separated due to the movement of the finite verb part to C
(V2). Compare with the uninflected, unmoved infinitival counterpart in (33), which has the
same pragmatics as the inflected imperative in (32).
(32)
Bass auf!
Watch out
‘Be careful!’imperative
(33)
Aufbass’n!
Out.watch
‘Be careful!’infinitive
(34) shows that fei is compatible with the imperative, in fact this is a use of the
particle that is widely attested in BG. (35) shows that it is impossible to combine the
particle with an infinitival command, presumably due to the lack of CP structure.
(34)
a. Bass fei auf!
Watch out
‘Be careful!’imperative
(35)
*fei aufbass’n
fei out.watch
*infinitive
In this section I showed that the proposed analysis of fei makes some predictions
that are borne out by the data. Fei, hosted in the CP layer of the clause, is incompatible with
infinitival commands, which lack that layer. In addition, the particle is ungrammatical in
questions, nor can it be used in an answer to a y/n question. I took this to be due to an
incompatibility between the illocution type of questions and the assertion expressed with
fei.
5. A brief outlook: fei and doch
Fei as described in this paper has many similarities with another particle found in
SG and BG, namely doch9. Doch shares a similar semantics with fei, and can be
felicitously uttered in many of the fei contexts. According to Zimmermann (2007b) doch(p)
“is used to express the speaker’s assumption that the addressee is not aware of p. For
instance, the addressee may have (temporarily) forgotten about p, or she may think p false”.
9
I only address unstressed doch here. The difference between stressed and unstressed particles is hotly
debated, and a closer investigation is currently beyond the scope of this paper. Note that fei can never be
stressed, unlike most discourse particles.
At a first glance, this description of doch is very similar to the proposed meaning of fei: ‘ I
belive that you believe that ¬p, but p!’ On a closer look, however, a crucial difference is in
the assumed knowledge state of the addressee: fei is felicitous in cases where the addressee
actively entertains ¬p (or the speaker assumes that the addressee actively entertains ¬p);
doch is felicitous where the addressee is (currently) unaware of p.
Consider again the OOB example (2), here as (36), with an infelicitous doch.
(36)a. #Dei Hos’ntiarl is doch auf
#your pant.door is doch open
To mean:‘Your fly is down’
b. Dei Hos’ntiarl is fei auf
your pant.door is fei open
‘Your fly is down’
Doch is infelicitous in (37a), since the speaker implies with its use that the addressee at
some earlier point was aware of the fact that their zipper is down, but has forgotten it. This
interpretation is odd, resulting in the sentence being bad.
In a case that can be construed either way, i.e. as the addressee being temporarily
unaware of p OR actively entertaining ¬p, both doch and fei are good (23 repeated as 37).
(37)
A:
Can I open the window?’
a. B:
Mi friats
Me is.cold
‘I am cold’
fei
fei
‘You seem to believe that I am not cold, but I am cold’
b. Mi friats doch
Me is.cold doch
‘I am cold’
‘ Remember, I am cold’
Due to fei, (37a) has to be construed as A actively entertaining that B is NOT cold.
Example (37b) has to be interpreted as A knowing that B is cold, but having forgotten.
Both are possible and good.
In sum, both fei and doch are contrastive in meaning, but a closer look reveals that
each particle requires a subtly different requirement on the knowledge state of the
addressee10.
1. Summary
I argued that the BG particle fei encodes polarity focus by contrasting the
alternatives p and ¬p, and also acts like a discourse particle, since it comments on
presuppositions in discourse and corrects them. I proposed that fei(p) expresses the
speaker’s assumption that the addressee is actively entertaining ¬p and instructs the
addressee to accept p as part of the CG: ”I am asserting p, since judging from your
behaviour/question/comment you seem to think ¬p, or just in case that you think ¬p”.
10
An in-depth discussion of the similarities and differences of fei and doch is not at the core of this paper, and will
have to be left for further research.
References
Abraham, W. ed. 1991. Discourse Particles: Descriptive and Theoretical Investigations on the Logical,
Syntactic and Pragmatic Properties of Discourse Particles in German.Amsterdam: Benjamins.
Bayer, J. 2001. Asymmetry in emphatic topicalization, In: C. Féry & W. Sternefeld (eds.) Audiatur Vox
Sapientiae. Akademie-Verlag, Berlin. 15-47.
Den Besten, H. 1983. ‘On the Interaction of Root Transformations and Lexical Deletive Rules.’ in Werner
Abraham (ed.) On the Formal Syntax of the Westgermania. Papers from the “3rd Groningen
Grammar Talks” Groningen, January 1981. Amsterdam/Philadelphia: Benjamins. Linguistik Aktuell
3:4.
Fischer, K. ed. 2006. Approaches to Discourse Particles. Elsevier.
Hamblin, C. L. 1973. Questions in Montague English. Foundations of Language 10: 41-53.
Höhle, T. 1982. Explikation für „normale Betonung“ und „normale Wortstellung“. In W. Abraham (ed.):
Satzglieder im Deutschen: Vorschläge zur syntaktischen, semantischen und pragmatischen Fundierung.
Tübingen: Narr. 75–153.
Höhle, T. 1992. 'Über Verum-Fokus im Deutschen', in J. Jacobs (Ed.), Informationsstruktur und Grammatik,
Opladen, Westdeutscher Verlag, 112-141.
Jackendoff, R. 1972. Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT Press.
Jacobs, J. 1991. "On the semantics of modal particles", in: Abraham, Werner (ed.): Discourse Particles.
Amsterdam / Philadelphia: Benjamins, 141-162.
Jacobs, J. 1992. "Neutral Stress and the Position of Heads", in: Jacobs, Joachim (ed.): Informationsstruktur
und Grammatik. Opladen: Westdeutscher Verlag, 220-244.
Kadmon, N., 2000. Formal Pragmatics Blackwell, Oxford.
König, E. 1991. The meaning of focus particles. A comparative perspective. London, New York: Routledge.
Krifka, M. 1998. Additive particles under stress. In Proceedings of SALT 8, Cornell, 1999. CLC Publications,
111–128
Romero, M. and C.-H. Han. 2001. On Certain Epistemic Implicatures in Yes/No Questions. Proceedings of
the Thirteenth Amsterdam Colloquium, ed. By Robert van Rooy. Amsterdam: ILLC/Department of
Philosophy, University of Amsterdam. Pp.168-173.
Rooth, M. 1996. Focus. In: Lappin, S. (ed.) Handbook of Contemporary Semantic Theory. Oxford: Blackwell,
271-297.
Weydt, H. 1969. "Vorwort". In: Weydt, Harald (ed.): Aspekte der Modalpartikeln: Studien zur dt. Abtönung.
Tübingen: Niemeyer
Zimmermann, M. 2004. Discourse Particles in the Left Periphery", in: Shaer, Benjamin / Frey, Werner
/Maienborn, Claudia (eds.): Proceedings of the Dislocated Elements Workshop, ZAS Berlin, November
2003. ZAS Papers in Linguistics 35, 543-566.
Zimmermann, M. 2007a. Contrastive focus. In Féry, C., G. Fanselow, and M. Krifka (eds). Interdisciplinary
Studies on Information Structure 6 (2007) 147-159.
Zimmermann, 2007b. Discourse Particles. To appear in P. Portner, C. Maienborn und K. von Heusinger
(eds.), Handbook of Semantics. (= Handbücher zur Sprach- und Kommunikationswissenschaft). Berlin,
Mouton de Gruyter.
Sonja Thoma
Dept. of Linguistics, University of British Columbia
[email protected]