Professional Documents
Culture Documents
__________________________________________
John T. Beavers, Ph.D.
Department of Linguistics
Supervising Professor
__________________________________________
Hans C. Boas, Ph.D.
Department of Germanic Studies
Second Reader
1. ABSTRACT
Author: Amanda Haight
Title: Beyond Case Marking in the Interpretation of German Prepositional Phrases
Supervisor: John Beavers
Languages employ a variety of methods for distinguishing directed motion from locational
motion. Regardless of encoding method, many languages exhibit a pragmatic reinterpretation
process where the locational construction is coerced by a variety of factors, including verb class,
landmark, syntactic particle, and other contextual material to mean the goal of directed motion.
Given the diverse nature of the languages employing this pragmatic tool, it was hypothesized
that pragmatic directional reinterpretation of locations was a universal feature of languages
(Beavers et al. 2010, Tham et al. 2012).
After examining an online corpus and consulting on constructed, controlled examples with
a native speaker of German, this thesis concludes that this pragmatic reinterpretation is not
available in Standard German. Instead of German locations being reinterpreted as goals of
motion in contexts necessitating a directional reading, explicit directional encoding becomes
obligatory and locational encoding becomes unacceptable instead of reinterpreted. This finding
calls into question the universal nature of the hypothesis set forth by Beavers et al. (2010) and
Tham et al. (2012), suggesting parametric variation of pragmatic directional reinterpretation of
locations across languages.
TABLE OF CONTENTS
1.
ABSTRACT
........................................................................................................................................
i
2.
INTRODUCTION
.............................................................................................................................
1
3.
LITERATURE
REVIEW
..................................................................................................................
6
3.1
Talmys
Typology
......................................................................................................................................................
6
3.2
Effects
of
Landmark
and
Ground
.......................................................................................................................
7
3.2.1
Areas
v.
Container
...........................................................................................................................................
7
3.2.2
True
Covers
v.
Cover
Proxy
.........................................................................................................................
8
3.2.3
Tightness
of
Landmark
..................................................................................................................................
9
3.3
Effects
of
the
Verb
.................................................................................................................................................
13
3.3.1
Verb
Classes
....................................................................................................................................................
13
3.3.2
Duration
and
Punctuality
..........................................................................................................................
16
3.4
Directional
Particles
.............................................................................................................................................
17
3.5
Effects
of
Syntactic
Arrangement
...................................................................................................................
18
3.6
Proximity
of
Figure
to
Landmark
...................................................................................................................
19
3.7
Evidence
of
Reinterpretation
Across
Typologically
Distinct
Languages
.......................................
20
4.
CORPUS
STUDY
...........................................................................................................................
22
4.1
Goals
of
Study
..........................................................................................................................................................
22
4.2
Optimality
Theory
Guided
Case
Choice
........................................................................................................
22
4.3
Methodology
of
Corpus
Study
..........................................................................................................................
24
4.4
Results
of
Corpus
Study
......................................................................................................................................
25
5.
NATIVE
SPEAKER
CONSULTATION
......................................................................................
27
5.1
Methodology
of
Native
Speaker
Consultation
...........................................................................................
27
5.2
Irrelevant
Factors
..................................................................................................................................................
28
5.3
Motion-Landmark
Conflicts
..............................................................................................................................
30
5.3.1
Motion
Does
Not
Conflict
With
Landmark
.........................................................................................
31
5.3.2
Motion
Conflicts
with
Landmark
...........................................................................................................
33
5.3.3
Motion
Conflicts
with
Source
...................................................................................................................
34
5.4
Effects
of
Landmark
..............................................................................................................................................
35
5.4.1
Areas
v.
Containers
......................................................................................................................................
35
5.4.2
Nested
Landmarks
.......................................................................................................................................
42
5.5
Effects
of
Verb
Classifications
..........................................................................................................................
43
5.5.1
Directed
Motion
Verbs
...............................................................................................................................
43
5.5.2
The
Clothesline
Problem
...........................................................................................................................
46
5.5.3
Verb
Summary
...............................................................................................................................................
49
5.6
Effects
of
Syntactic
Elements
............................................................................................................................
50
5.6.1
Directional
Particles
....................................................................................................................................
50
5.6.2
The
Bridge
Problem
(Alternation
in
Path
Encoding)
....................................................................
53
6.
CONCLUSION
................................................................................................................................
57
7.
BIBLIOGRAPHY
...........................................................................................................................
60
8.
BIOGRAPHY
.................................................................................................................................
62
ii
2. INTRODUCTION
German is a case marking language, which means the role of a noun phrase in a clause is
indicated through morphological marking within the noun phrase itself.1 Prepositional phrases
(PPs) are a special circumstance when it comes to case marking because they can determine
much more about the noun than simply whether the noun phrase is a direct or indirect object,
which the accusative and dative case mark in simpler German sentences. Instead, PPs can
determine the semantic role a constituent plays in the sentence, such as whether it is the route,
source, instrument, or what I will be focusing on: goal and location. 2
German prepositions are divided into four classes by the case the constituent noun phrase
(NP) is assigned: dative, accusative, genitive, and alternating.3 The constituent NPs of this latter
set of alternating prepositions can be marked by either the dative or the accusative depending on
the semantic intentions of the speaker. The traditional grammar-school rule goes something like
move-ative accusative, state-ive dative; that is to say the accusative indicates that the NP is the
goal of motion in a description of directed motion, and the dative marks the location where
motion is occurring (Haider 2010: 241). A simple actualization of this rule is seen in (1).
Special acknowledgement goes to Santiago Sanchez who helped build the script that searched
the German corpus for examples relevant to this study, and to Katrin Fuchs who generously
2
This study is concerned strictly with Standard German. In non-standard German, such as
regional dialects, case assignment following prepositions functions significantly differently. All
assumptions and data henceforth are in reference to acceptability in Standard German.
3
Dative - aus, out of, auer besides, bei by or at, mit with, nach after, seit since, von
from, zu to; accusative - bis until, um around, fr for, durch through, ohne without,
gegen against, contrary to; genitive - anstatt instead of, innerhalb inside, auerhalb
outside, trotz in spite of, whrend during, wegen because of; alternating - an on (vertical
surface), auf on (horizontal surface), hinter behind, vor before, in front of, ber over,
unter under, in in, neben next to, zwischen between
1
Whereas German has explicit case marking to determine the goal or location implications of
prepositions, English, excluding into and onto, marks this alternation through context instead of
markings. To explicitly mark the location reading of in, on and other prepositions in English
translations, I will use inL, onL, etc. which can conventionally be read as around in, around on.
Wholly ambiguous prepositions will be marked with L for readings of location and G for readings
of goal.
5
Notably excluding to, which has a strictly dynamic goal reading.
2
In English, into is strictly dynamic. However, in is not strictly stative, meaning there are contexts
in which the location variant in is acceptable with an explicitly dynamic reading. For example,
the use of inG in (4) can be replaced with into, but the use of the nominally stative preposition is
still acceptable on this reading.
(4)
Similarly, there are noted examples in Russian where the locative marker is acceptable in
describing an endpoint of motion, as seen in (5).
(5)
Beyond English and Russian, there has been similar variation noted in Polish (Cienki 1989:1417), Czech (Belichova-Krzhizhkova 1974; Ungermanov 2005), and Ukrainian (Nedashkivska
2001). In (6)-(8), the French preposition dans is a typically locative preposition, but there are
certain contexts and elements that contribute to a goal reading.
(6)
(7)
(8)
Drawing from typologically distinct satellite-framed and verb-framed languages (satellite- and
verb-framed languages to be discussed further in section 3.1) Beavers et al (2010) suggest
directional interpretations of locative adpositions should be available with the appropriate
pragmatic support even in the absence of morphosyntactic devices for directly expressing
3
direction in a PP, (Beavers et al 2010 : 362). This hypothesis is further elaborated in Tham et al
(2012): The process that derives directional readings is available across languages, even in
language types that differ from English (Tham et al 2012). These researchers consistently
propose that locative PPs are reinterpretable as directional PPs in specific pragmatic contexts
across language typologies. These pragmatic contexts vary across languages, but include factors
such as the type of ground or landmark, the verb class (durative v. punctual, change of state v.
change of position, etc.), permanence of the resulting new location, syntactic arrangement, and
the presence of directional particles.
If the conclusion from Tham et al (2012) Beavers et al (2010) is correct, then it is an
attested cross-linguistically available option to encode directed motion using a locational PP in
combination with specific pragmatic contexts, and furthermore this option seems to exist in a
typologically diverse set of languages. From this we might infer that it is a universally available
process. In this thesis I examine the case study of German, a language with an alternation
between explicitly locational and directional encoding, as in Russian. The null hypothesis is that
there should be some contexts where the location (marked by the dative case) can be
pragmatically reinterpreted as the goal of motion (typically marked by the accusative case) in the
same or similar contexts in which this is attested in other languages, contrary to the standard
grammar school rules.
I discuss evidence from corpus data and native speaker intuitions controlling for the
contexts and relevant factors noted for other languages, and show that a divergence from this
alternation was not attested in German. In other words, this study strongly suggests that locations
categorically cannot be reinterpreted as goals in German, and that the universality of pragmatic
reinterpretation as asserted by Tham et al (2012) and Beavers et al (2010) is not valid.
This suggests that the presence of this strategy in a language is itself subject to
independent parametric variation, much like the various factors Beavers et al (2010) suggest
cross-cut other languages, albeit here having to do with a pragmatically conditioned interpretive
process rather than mophosyntactic processes.
To reach this conclusion, I will first review in section 3 the relevant literature attesting to
divergence in the alternation of goal and location marking in languages other than German,
noting specifically the effects of the landmark, the verb, directional particles, and syntactic
arrangement. Section 4 begins the search for divergence from the standard German
generalization by extracting and examining literal uses of unter in the sDeWaC corpus.6 Section
5 details the results found after consulting a native speaker about case acceptability and
subsequent semantic interpretation of novel examples constructed with guidance from
divergence found in other languages. While the factors tested in section 5 clearly affected case
acceptability and semantic interpretation, there was no evidence for reinterpretation of locations
as goals in German. Section 6 discusses the theoretical importance these findings implicate and
suggests further investigation as to why certain languages allow this type of pragmatic
reinterpretation and others do not.
3. LITERATURE REVIEW
Before discussing German data, I briefly recap relevant background information on the
expression of directed motion across languages, starting with Leonard Talmys typology of
motion expressions, moving to a discussion of contexts outside of German where overtly marked
locations are pragmatically reinterpreted to be the goals of motion.
The examples of Spanish (9a) and English (9b) are prototypical examples of verb-framed
and satellite-framed languages respectively, though it is important to note that this typology is
best interpreted as a tendency rather than a necessity, e.g. satellite-framed languages often have
directionally encoded verbs like the English enter (Beavers et al 2010).
(9) a. Entr corriendo/volando/nadando a la cueva.
enter.pst.3sg running/flying/swimming a the cave
He entered the cave running/flying/swimming. (Spanish Talmy 1985:111)
b. He ran/flew/swam into the cave.
It is this typology and understanding of German as a satellite-framed language that forms the
foundation to explore the directionality of motion events described by German prepositional
phrases. In German, directionality can be determined in a number of ways, but most often
through a satellite PP and the case on its complement NP.
Areas v. Container
Nikitina divides landmarks into areas and containers and asserts that inG is more likely to
occur with landmarks that have the qualities of a container. Nikitina distinguishes these
landmarks by the presence (containers) or absence (areas) of well-defined boundaries. In her
7
corpus, Nikitina found that inG was used with 9% of area landmarks and 17% of container
landmarks. While into was always the preferred method for conveying directed movement, there
was a statistically significant difference in the use of directional inG between containers and
areas. She accounts for this difference in frequency by noting that areas are more likely to be
associated with motion with prominent paths, thus using in is more likely to be interpreted as a
locational PP and into is used to explicate the directionality of movement. Thus, because the
landmarks lack distinct boundaries, Nikitina argues that examples like (10b) are more likely to
be interpreted as locational than directional and thus are more likely to employ into to explicate
directed movement, unlike (10a).
(10) a. He walked in the room/backyard/store. (locational or directional)
b. He walked in the city/field/mountains. (locational / ??directional)
(Nikitina 2008 : 187 (19))
For our study, I will test the effect of areas and containers on acceptable case marking to
determine if the dative can ever be reinterpreted as marking a goal of directed motion. From
Nikitinas study, I would expect to find a comparable directional interpretation of a dative
marked PP constituent meaning inG when the motion involves a container landmark.
3.2.2
Where containers and areas discussed by Nikitina (2008) pertain specifically to the preposition
in, Estigarribia and Levin (2007) discuss similar constraints on landmarks of under, notably
distinguishing true cover and cover proxies. A true cover seen in (11) is a structure with a clear
enclosed space underneath. It is analogous to a container and more readily receives a directional
reading of under. A cover proxy seen in (12) is a larger entity without ground support. It is
analogous to areas and more readily receive a locational reading of under.
Tightness of Landmark
This section discusses a noted variation in Russian where an implied boundary crossing is
marked by a locational marker. Israeli (2004) claims there are certain contexts that allow
speakers of Russian to emphasize the motion within a landmark through the use of a locational
marker rather than emphasize the motion across a boundary through the use of a directional
marker. These constraints center around the size of the figure in comparison to the size of the
landmark and how snuggly the figure fits within the landmark and any other occupants of that
space.
In Russian placement verbs the accusative case, like in German, is used to mark the goal of
motion, and either the locative or instrumental case can be used in a stationary event. This
standard generalization for Russian is illustrated by (13a,b).
(13) a. Malik sel na divanA.
boy.NOM.SG sat.PF.M on couch.ACC.SG
The boy sat on the couch.
b. Malik sidel na divaneL.
boy.NOM.SG sat.IPS.M on couch.LOC.SG
The boy was sitting on the couch.
(Israeli 2004)
In concurrence with the standard pragmatic reinterpretation hypothesis and other literature on
directed motion in Russian, Israeli asserts that there is evidence of objects with locational
marking begin reinterpreted as goals of motion given appropriate context or intention of the
9
speaker. An example of this variation in Russian is seen in (14) where word order and emphasis
affect the speakers choice of accusative or locative to mark the landmark, but semantics is the
same.
(14) a. Poves kartinu na stenuA.
Hang.IMP.SG painting.ACC.SG on wall.ACC.SG
Hang the picture on the wall.
b. A na steneL my povesim kartinu.
And on wall.LOC.SG we.NOM hang.1P.FUT.PL painting.ACC.SG
And on the wall we will hang a picture.
(Russian Zaitseva 1994: 112)
Israeli uses the schemas in Figure 1 to illustrate case choice in these placement verbs. The source
of variation comes particularly from the speakers choice between figures 1.3 and 1.4.
Figure 1 Image Schemas, Israeli (2004:9)
1.1
1.2
1.3
1.4
LOC/INST: No motion
LOC/INST: Path of motion does not cross boundary
ACC: Path of motion crosses boundary
LOC/INST: Path of motion crosses boundary, emphasis on motion inside container
Israeli differentiates figure 1.4 from 1.3 by claiming that though the crossing of a boundary is
implied by the description, it is the motion inside the container that is emphasized and triggers
the otherwise stationary case. Israeli cites Russian examples that clearly imply the crossing of a
boundary, yet are marked by the locative case, as in (15). In this case, the container must be large
enough to contain multiple of the moving objects in question, yet said object when coming to
rest inside the container not snuggly fit (see figure 2.4 below). While (17) and (18) seem
contradictory, Israeli attributes the conflict to a matter of emphasis of motion.
10
(15)
(16)
(Israeli 2004)
Israeli claims that the use of the locative in (15) emphasizes the motion within the container,
rather than the inherent crossing of the container boundary. In (16) the boundary crossing is
emphasized by the use of the accusative, though the ground is also a large container.
Nonetheless, the data show that the locative case can be used to encode a goal.
That said, only certain contexts are subject to a choice of emphasizing the boundary
crossing or the motion within the container, affected primarily by the size of the figure compared
to the size of the landmark. If the figure fits snuggly in the landmark, the accusative must mark
the goal. Only when the figure fits loosely in the landmark is the locative or instrumental case
an option. This fitting model is illustrated by figure 2.
Figure 2 Israelis container illustrations
Israeli notes here that the context of this example is a large garage, accommodating more than
one car, that is relatively empty. A car parked in a single car garage would fit the structure of 2.1
and be marked by the accusative.
8
Glossing conventions here borrowed from the source.
11
Israeli notes that objects that fit singularly or snuggly inside a container (a body in a coffin,
a person in a chair, a letter in a pocket, etc., Israeli 2004:16) are regularly are marked by the
accusative case in concomitance with Figure 1.3. The car in the garage in (15) fits loosely, and is
therefore subject to emphasis of the movement within the landmark. Example (17) illustrates the
new case acceptability if the garage discussed in (15) was filled with multiple cars, and the
ground would be marked by the accusative and fit with figure 2.3.
(17)
Through Figures 1 and 2, Israeli illustrates the relation between the figure and the ground as
pivotal in case choice in Russian placement verbs. In particular, there are certain constraints that
allow a subset of boundary crossings to be described by locational marking. If the figure fits
loosely within the landmark, allowing for emphasizeable movement within the landmark, the
speaker can choose to emphasize this interior movement rather than the boundary crossing
through the choice of locational marking.
Israeli touches on, but does not attempt to address, a second set of variations from
Russians standard generalization. Whereas the above mentioned examples allow for both the
accusative or locative/instrumental markings depending on contextual constraints and speaker
emphasis, there is a second category of variation where not only is locational marking used to
mark a goal, but the accusative is entirely unacceptable. One such example is seen in (18).
(18) a. *On povesil rubaku vo dvorA.
he.NOM hung.PF.TR.M shirt.ACC.SG in courtyard.ACC.SG
He hung his shirt in the yard.
12
Verb Classes
Nikitina breaks down Russian motion verbs into the following classes by their reading and
marking acceptability:
General motion entailment verbs: can combine with either a location-marking or goalmarking satellite to describe location or goal of motion respectively
Inherently undirected motion verbs: disallow goal marking satellites
Inherently directed motion verbs: disallow location marking satellites
Encoding variation verbs: both location and goal satellites mark the endpoint of motion
It is this last category that is of most interest to this study. The verbs that make up this last class
are change of position (saat seat (caus.), klast lay/put/place down, stavit put/place/set,
veat hang (caus.) and their intransitive counterparts), change of state verbs (prjatat hide,
zatoat incarcerate, zaryvat bury, etc.), and less importantly metaphorical motion verbs
(zapisyvat to write down) where both locational and directional PPs can express the goal of
13
motion. The alternation with what Nikitina calls change of position verbs is exemplified in
English through (19).
(19) a. John put the box into his pocket.
b. John put the box in his pocket
Comparing (19) with (20) specifies that this alternation is based in the verb and not the relation
of the figure to the landmark as discussed in section 3.2.3. Essentially, put in English requires a
directional reading, but does not require a directional marking.
(20) a. John put the ping-pong ball into the large empty box.
b. John put the ping-pong ball inG/*inL the large empty box.
A similar variation occurs in Russian, as seen in (21), as well as Polish, Czech, and Ukrainian.
Both the locational and directional markings are acceptable. A similar alternation is illustrated in
(22) with change of state verbs.
(21)
Change of position
Postav vazu na stol/stole.
Put vase on table.ACC/table.LOC.
Put the vase on the table.
(22)
Change of State
Pirat sprjatal zoloto v sunduk / v sunduke.
Pirate hid gold in chest.ACC / in chest.LOC
The pirate hid the gold in a chest.
Examples (19)-(22) arguably involve change in location since there is no explicit contextual
source, but it could be implied which would contribute to the directional reading. But the explicit
lack of change in location in (23) explicates that there is no inherent change of location
necessarily entailed by change of position and change of state verbs.
(23) a. Ja uvidel, to butylka upala, I snova e postavil.
I saw that bottle fell and again it.ACC put.up.
I saw that the bottle fell and put it back into standing position.
14
Though the acceptability alternation is not attested by change of position verbs in German,
Nikitinas suggestion that the use of the accusative actually indicates a result argument rather
than a goal argument is significant in the interpretation of the uniquely acceptable case in
German. While German does not exhibit the same acceptability alternation, a similar semantic
implication may be possible.
15
3.3.2
Duration and punctuality of a verb is noted by Nikitina (2008) to influence the choice of a
locational or directional marker in English. Specifically, transition events (punctual verbs) are
more likely than process events (durative verbs) to allow a directional in. While into is always
preferred in English when describing a directional event, Nikitina attests that in is sufficient in
transition events as a directional reading and is more likely to depict a directional motion than
with process events. When describing directionality, process events are more likely to occur with
into to explicate directionality as in can be more easily interpreted in this class of events as
locational than directional.
Nikitina concludes that verbs that tend to describe transitions (such as get, bring, put, dip)
occur more frequently with directional in than verbs that describe inherently unbounded
processes (such as climb, drive, pull, carry) (Nikitina 2008:181). That is to say, with verbs that
describe transitions, the ambiguous in is often read as the goal of the transition, and with verbs
that describe processes, the ambiguous in is often read as the location of the process.
This assertion is backed up by Nikitinas data, which found that a directional interpretation
of in (inG) was used in 19% of cases with transition verbs and 12% with process verbs, a
statistically significant difference. In either case, into is preferred to inG when talking about
directed motion, and the difference between the 19% and 12% is only 3 clauses of the dataset (37
instances of inG with process verbs compared to 40 instances of inG with transition verbs).
This study highlights the relationship between a quality of the verb and the acceptability of
reinterpreting a location as a goal. In English, both durative and punctual verbs allow such a
reinterpretation, however it is more likely to occur during transition events with punctual verbs.
16
example shows the assertion that directional particles contribute to the directional reading of a
locative marking is valid. The examples in (27) were constructed with the same unbounded verb,
but with a non-conflicting landmark (the trees). In contrast to the streetcar, the trees is a
landmark that can be climbed in an unbounded manner, la Tarzan. Example (27a) maintains a
primarily locational reading, compared to (26b) which has a directional reading. Examples
(27b,c) illustrate that a directional reading can be induced through the use of a directional
satellite or a directional particle.
(27) a.
b.
c.
d.
(location, ??direction)
(direction, *location)
(location, direction)
(direction, *location)
Though the examples provided by Nikitina did not strongly support the assertion as they
were compounded by other directional influences, a reconstruction controlling for other
directional influences reveals that the addition of directional particles specifies the directional
contexts, rendering the directional satellite unnecessary to attain a directional reading. Through
the use of a directional particle in English, a location can be reinterpreted as the goal of
movement.
18
Russian allows for a fairly free rearrangement of constituent elements, as case aids in
interpreting argument roles. While the traditional syntactic arrangement illustrated by (29b)
allows for an alternation in interpretation, the preposing of the PP in (29a) disprefers the
accusative marking.
(29) a. V kuxne na ?9stenu / na stene ja povesil asy.
In kitchen.LOC on ?wall.ACC/on wall.LOC I hang clock.
In the kitchen on the wall I hang a clock.
b. Jam povesil asy v kuxne na stenu / na stene.
I hang clock in kitchen.LOC on wall.ACC/on wall.LOC.
I hang the clock in the kitchen on the wall. (Russian Blazhev (1988:64-5))
The comparison of (29a) to (29b) indicates that topicalizing the location in (29a) prefers the use
of locational PPs. In other words, (29a) emphasizes the place where the clock is being hung and
(29b) emphasizes the change of location of the clock, while both imply motion and change of
location of the clock.
The conclusions from these studies indicate that syntactic arrangement, particularly
preposing the alternating argument in question, can make an ordinarily alternating PP show
preference for a particular marking or reading. In practice, this means this study will need to
control for syntactic arrangement lest the variation is missed.
Here and below, a single question mark identifies another possible marking that retains the
same semantic interpretation, but indicates the dispreference for this form compared to the
alternative.
19
directional reading is more available for in the closer the figure is to the landmark. The locational
in in (30a) receives a directional reading more easily than in (30b) because the figure is closer to
the landmark.
(30) a. Pat walked inG the study from the kitchen.
b. #Pat walked inG the study from a mile away.
When considering border-crossing events, it will be pertinent to examine the scalable proximity
of the figure to the landmark and its effect on case acceptability or preference in German.
20
(32)
Even languages outside of the standard verb-framed and satellite-framed languages show
evidence for directional reinterpretation of locative markings. Mandarin, an equipollently-framed
language, uses coverbs, lexical units exhibiting properties of both verbs and prepositions (Li and
Thompson 1981), to express location and directionality.10 The coverb zi in isolation is read
exclusively as locative, in contrast with the directional coverbs do arrive/to and jn enter.
However, given similar contexts that govern reinterpretation in S- and V-framed languages, zi
can receive a directional reading, as seen in (33)-(34).
(33)
Dio zi shu-li
Drop be.at water-within
drop into the water
(34)
While the variation is not identical across languages (the exact verbs contributing to a directional
reading in one language do not always correspond exactly in another language), there are
consistent patterns suggesting that certain types of verbs, like displacement, transition, and
arriving verbs, and types of landmarks contribute to the directional reinterpretation of nominally
locative markings along with other contextual patterns. Guided by the principal that there are
cross-linguistic patterns contributing to directional reinterpretations, the reinterpretation
hypothesis suggests that German should allow this reinterpretation given similar contexts.
10
In equipollently-framed languages, both manner and path are both encoded as main verbs
(Beavers et al. 2010).
21
4. CORPUS STUDY
4.1 Goals of Study
In quick review, the traditional understanding of the use of either accusative or dative case in
German to mark the complement NP of a motion prepositional phrase is that accusative marks
the NP as the goal of the motion and dative marks the NP as simply the location of the motion.
As evidenced by the previously discussed literature, in languages that exhibit a similar
goal/location dynamic, there are certain situations where this is complicated or there are
exceptions to this rule. Specifically, I am looking to either corroborate or refute the interpretation
of Beavers et al (2010) and Tham et al (2012) that languages universally can reinterpret locations
to be goals under certain pragmatic conditions. In the context of German, I am looking for data
where the dative is used to mark a NP that is understood to be the goal of motion.11
11
its complement must always take the dative case. This demonstrates that dative case can be
associated with a goal reading if it is lexically fixed or the only possible complement case for a
given preposition. Zwarts (2006) discusses a hierarchy of constraints in Optimality Theory that
accounts for case choice with non-alternating prepositions, placing lexicalized case above
contextually determined case. The question I address here is not whether dative can in general
ever encode a goal, but rather whether it does when it alternates with accusative.
While this study was primarily concerned with alternating prepositions, one tested example
explored the possibility of combining directional selecting verbs (stellen, legen, and setzen) and a
strictly dative preposition, bei by, at. For this quick study, setzen to sit was left out as it
required an object with an anatomical seat. The examples in (35b) and (36b) exhibit the
standard generalization when an accusative preposition is used. A native speaker was unsure as
to which case was acceptable in (35a), suggesting that an unacceptability lay in an improper
combination of verb and preposition. But it is the result of (36a) which was quite unexpected.
While there was a preference to use an alternative preposition to communicate this event, the
native speakers intuition was that the accusative, not the dative case was acceptable, and she
suggested that (36a) might be a colloquialism, or something a rural grandmother might say.
(35) Legen to lay sth.
a. Ich lege das Buch bei *den/*dem Baum.
I lay the book by the.ACC/the.DAT tree.
I lay the book by the tree.
b. Ich lege das Buch neben den/*dem Baum.
I lay the book next to the.ACC/*the.DAT tree.
I lay the book next to the tree.
(36) Stellen to put sth.
a. Ich stelle das Buch bei den/*dem Baum.
I put the book by the.ACC/*the.DAT tree.
I put the book by the tree.
23
12
The following are the strategically chosen verbs: laufen to walk, fahren to drive, steigen to
climb, werfen to throw, fliegen to fly, rollen to roll, wandern to wander, gleiten to glide,
schwimmen to swim, tanzen to dance, krabbeln to crawl, rennen to run, hpfen to
jump/bound, schlngeln to wiggle, streifen to roam, wackeln to totter, radeln to bike,
24
an element of a separable prefix or particle verb, it is likely to be used in describing literal, not
metaphorical, motion, and it is predicted that it will be easy to determine the source, route, and
goal of the given syntactic event.
From this set of data I manually reviewed the sentences and further extracted only
examples where unter was a dependent of the verb in question. Further complicated by Germans
widespread use of particle verbs, which often use prepositions as a non-literal complement to
invoke a semantic meaning distinct from that of the main verb, these examples were further
narrowed to keep only the examples where unter was used in a literal sense of referring to an
event happening under some physical object, not say, under the supervision/gaze/cheers of the
overseer/neighbors/crowd.
bummeln to stroll, joggen to job, hinken to limp, traben to trot, humpeln to hobble,
hoppeln to scuttle, hopsen to hop, rudern to row, trieben to float.
25
19% locational readings of under. Having tested similar constraints on the German corpus, it is
extremely surprising to have found absolutely no alternation or divergence in German case
marking.
While the absence of divergence is striking given the breadth of the corpus study, it could
be attributed to a number of assumptions and constraints aimed at limiting the data. It could be
that an alternation is not available for unter but is for in, or there is another context outside of
verb class that allows for the directional interpretation of locational marking. The constraints
imposed to make the data manageable theoretically could have eliminated the sought after
exceptions. Thus a second methodology of constructing examples was used to further examine
the expanded contexts in which a deviation from the standard generalization might occur in a
more systematic and controlled manner.
26
13
Examples in the following section may employ one or both of these context presentation
strategies as the data permit, but the strategy of context presentation proved to have no affect on
acceptability and should not be considered a testable factor.
27
inG or into) with verbs encoding distinct manner of motion (Nikitina 2008). The sentences shown
in (38) were constructed to test the effect of manner-heavy and manner-neutral verbs on
acceptability of directional or locational readings.
(38) a. Der Mann luft/fhrt/spaziert/springt in das/dem Zimmer.
The man walks/drives/wanders/jumps in the.ACC/the.DAT room.
The man walks/drives/wanders/jumps into/inL the room.
b. Der Mann luft/fhrt/spaziert/springt in den/dem Wald.
The man walks/drives/wanders/jumps in the.ACC/the.DAT forest.
The man walks/drives/wanders/jumps into/inL the forest.
While it was found that certain qualities of a verb do contribute to case acceptability (to be
discussed in section 5.5), manner was not one of them. The manner encoded in verbs did not
contribute to a divergence from the standard generalization, nor did it necessitate the use of a
particular case. Both cases were acceptable depending on the interpretation.
Syntactic Arrangement
As discussed in section 3.5, English and Russian showed a preference for locational readings
when the PP was topicalized and thus presented before the context. But German has a high
tolerance for topicalization while retaining meaning, and thus case preference is not affected by
syntactic arrangement. The comparison of (a-b) in examples (39)-(41) reveals that PP
topicalization does not change the acceptability of a given case nor the interpretation, and is thus
not a relevant factor when considering acceptability of cases in description of motion events.
(39) a. Der Mann spaziert in den/dem Wald.
The man wanders in the.ACC/the.DAT forest.
The man wanders into/inL the forest.
b. In den/dem Wald spaziert der Mann.
In the.ACC/the.DAT forest wanders the man.
Into/inL the forest wanders the man.
29
14
Given enough context focusing the motion within the room, the dative can be acceptable with
the interpretation of going back and forth inside the room, but the accusative use and intention
is consistently more likely in practice.
30
with a landmark, conflicts with a landmark, or conflicts with the source of motion and no
exterior context is presented.
5.3.1
When a motion can logically be accomplished within the bounds of the landmark, the use of the
dative is available to mark location and accusative is available for a directional interpretation.
This alternation of acceptable cases and subsequent interpretations is illustrated in (42). The
movements described by the verbs in (42) can reasonably be accomplished within the given
landmark, thus both case markings are available depending on semantic intention.
(42) a. Der Mann luft in das/dem Zimmer.
The man walks in the.ACC/the.DAT room.
The man walks into/inL the room.
b. Er fhrt in die/der Stadt.
He drives in the.fem.ACC/the.fem.DAT city.
He drives into/inL the city.
c. Er fhrt in die Berge/den Bergen
He drives in the.pl.ACC/the.pl.DAT mountains
He drives into/inL the mountains.
Walking can be accomplished in a room or can be a method of displacement into the room, thus
either case marking is acceptable for (42a) depending on the intention of the speaker. The
examples in (43) further illustrate possible case and interpretations when a movement does not
conflict with the given landmark.
(43) a. Der Mann luft in das/dem Haus.
The man walks in the.ACC/the.DAT house.
The man walks into/inL the house.
31
15
Verbs with high unintentionality like stroll, wander, or amble are unlikely to be used in the
context of crossing a boundary and thus use of the accusative would be rare and unusual, but
conceptually it is possible.
32
5.3.2
When motion conflicts with the landmark such that the specific motion within the landmark is
not possible, there is a preference for the accusative case because a locative interpretation is
either impossible or highly unlikely. This conflict of motion and landmark is best illustrated by
Israelis container illustration (see figure 2.1) where a movement is not logically possible given
the spatial configuration of the landmark. In section 3.4 I discussed the conflict between climb
and streetcar as the likely source of a directional interpretation in English. Climb in an
unbounded, continuous sense as demanded by a locational interpretation is logically improbable,
if not impossible, in the confines of a streetcar.
Example (44a) recreates this situation in German. Whereas English can reinterpret the
locative in as directional, German instead disallows the dative case and permits only the
accusative. This contrasts with (44b) in which the movement of walking does not inherently
conflict with the spatial configuration of the streetcar landmark. In comparing (44a) and (44b) it
is evident that it is not the presence of a directional particle (as was claimed by Nikitina 2008),
the unbounded nature of the verb (both steigen and laufen can be unbounded), or the size of the
figure that contributes to the locational marking acceptability. Rather it is the spatial requirement
of locational motion that conflicts with the spatial reality of the given landmark which allows
only a directional marking of (44a).
(44) a. Sie steigt in die/*der Straenbahn.
She climbs in the.F.ACC/*the.F.DAT streetcar.
She climbs into the/*inL the streetcar.
b. Sie luft in die/der Straenbahn.
She walks in the.F.ACC/the.F.DAT streetcar.
She walks into/ inL the streetcar.
A typically unbounded verb (like steigen to climb) is used to describe a bounded, transition
event in combination with a container-type landmark. However, laufen, also a typically
33
unbounded verb, is unaffected in case preference by the nature of the landmark. There is a
fundamental spatial requirement necessitated by a locative interpretation of steigen that is not
met by Straenbahn (i.e. steigen in the unbounded sense is physically not possible in the
confines of a streetcar). This suggests that there are spatial requirements for normally nondirected motion events that, when not met by the landmark necessarily yield a directional
reading, and in the case of German, require an explicit directional marking.
While this conflict of spatial requirements on motion and landmark spatial reality yields a
reinterpretation of locations as goals in languages other than German, this reinterpretation is not
available in this context in German and the conflicting marking is simply unacceptable rather
than reinterpreted.
5.3.3
In testing a set of contexts where the motion described conflicted with a source of motion, a new
subset of case marking constraints unattested by previous literature was discovered. Whereas a
conflict between motion and landmark yielded the accusative as the only acceptable marking,
(45) and (46) suggest that a conflict between the motion and the source yields the dative as the
only acceptable marking through similar logic.
(45)
(46)
16
The accusative variant is acceptable only when context specifies motion from one body of
water to another, as in rowing into the lake from a connecting river.
34
In (45), technically the swimmer is not swimming until they are actually in the water. Before the
jumping motion, they are standing, not swimming. In fact, it would not be possible to be on land
and have the moment of transition from cement to water to occur through swimming. A sense
of prototypical context accounts for the lack of explicit context for (46). Accusative would only
be acceptable if the rower was somehow in the boat on land, and through rowing actions, scooted
into the water and continued to row. In both examples, the type of motion described by the verb
conflicts with the source to elicit only a description of the post-transition motion and thus the
exclusive use of the dative case.
In comparison to the conflict discussed in section 5.3.2, this conflict derives from the
spatial requirements of the motion conflicting with the spatial realities of the source. Though
further data is needed to strengthen this hypothesis, (45) and (46) suggest that a conflict between
motion and source necessitates the dative as a description of only the post-transition event.
Though this conflict does not yield a reinterpretation diverging from the standard generalization,
it contributes to the understanding of case marking acceptability in German motion events.
Areas v. Containers
As was discussed in English (Nikitina 2008, Tham et al 2012) and Russian (Israeli 2004) in
section 3.2, the quality of the landmark affects the preference for a locative or directed
interpretation. As evidenced by the alternation with container landmarks in (47) and area
landmarks in (48), German shows no preference for locational or directional case marking.
35
(47)
(48)
In contrast to the increased English preference for directional in with containers (Nikitina 2008)
because of the defined boundary, German makes no distinction between the borders of areas or
containers. Rather the determination of a border-crossing event is determined by the choice of
case. Given no source context, the type of landmark, whether a container or area, was found to
have no preference for a locative or dynamic marking.
However, when context gave the source of motion indicating the presence or absence of a
border crossing, only the case marking corresponding to the directional (border crossing:
accusative) or locational (no border crossing: dative) event was acceptable. The examples in (49)
illustrate case choice in German where contextual evidence implicates the source of movement
as either on the inside of a border, the outside of a border, or on the border regardless of border
type.
(49) a. Non-Border Crossing
Sie steht in der Kche und dann luft sie in *das/dem Haus.17
She stands in the kitchen and then walks she in *the.ACC/the.DAT house.
She stands in the kitchen and then walks inL the house.
b. Border Crossing
Von dem Auto, luft sie in die/*der Bar.
From the.DAT car, walks she in the.ACC/*the.DAT bar.
From the car, she walks into the bar.
c. Source on the Border
Vom Waldrand, luft sie in den/??dem Wald
From the.DAT forest.border, walks in the.ACC/??the.DAT forest.
From the edge of the forest, she walks into/??inL the forest.
17
In context of traditional kitchen located inside the main building of house, not in separate
extension.
36
When the contextual source is inside the border of the landmark as in (49a), there is explicitly no
directionality and only the dative is acceptable. When the contextual source is outside the border
of the landmark as in (49b) there is explicit directionality and only the accusative is acceptable.
When the source of movement is explicitly on the border of the landmark, regardless of the type
of landmark, the accusative is always preferred. However, in this third context the dative could
be used to emphasize the movement within the landmark, though this expression would be rare
and somewhat unnatural. The use of the dative in this context is acceptable, but not preferred.
Examples (50) and (51) provide further evidence of exclusive dative acceptability in nonborder crossing events and accusative preference in on the border events.
(50)
37
(51)
In contrast to within the landmark and on the border motion, a border-crossing event involves
a scalable distance in terms of how close the figure is to the landmark. This proximity of the
figure to the landmark contributes to the allowance of a directional reinterpretation of in in
English the closer the figure is to the landmark (Thomas 2004, Tham et al 2012) as discussed in
section 3.6. Effect of proximity in the border-crossing event is tested in (52).
(52) a. CONTEXT : The woman is standing outside the entrance
Sie luft in die/*der Bar.
She walks in the.ACC/*the.DAT bar.
She walks into the bar.
b. Von gerade auerhalb des Eingangs, luft sie in die/*der Bar.
From right outside the.GEN entrance, walks she in the.ACC/*the.DAT bar.
From right outside the entrance, she walks into the bar.
38
c. CONTEXT : The woman is standing by her car in the parking lot, then enters
bar
Sie luft in die/*der Bar.
She walks in the.ACC/*the.DAT bar.
She walks into the bar.
Whether the woman in question (52) is close to or far from the border she crosses, the acceptable
case and reading is the same, indicating that in German, border-crossing events are also binary. If
there is a border-crossing event, no matter how far the figure is from the landmark, only the
accusative is acceptable. In German, proximity does not allow for the directional reinterpretation
of a location. Examples (53)-(59) are further evidence that proximity does not affect casemarking choice in border-crossing events.
(53) a. Er luft/hinkt/joggt/tanzt in das/*dem Zimmer von auerhalb des Hauses
He walks/limps/jogs/dances in the.ACC/*the.DAT room from outside
the.GEN house.
He walks/limps/jogs/dances into the room from outside the house.
b. CONTEXT : Man is standing outside of the house
Er luft/hinkt/joggt/tanzt in das/*dem Zimmer.
He walks/limps/jogs/dances in the.ACC/*the.DAT room from outside
the.GEN house.
He walks/limps/jogs/dances into the room from outside the house.
(54) a. CONTEXT : Man is standing outside of the room.
Er luft/hinkt/joggt/tanzt in das/*dem Zimmer
He walks/limps/jogs/dances in the.ACC/*the.DAT room.
He walks/limps/jogs/dances into the room.
b. Er luft/hinkt/joggt/tanzt von der Kche in das/*dem Wohnzimmer.
He walks/limps/jogs/dances from the.DAT kitchen in the.ACC/*the.DAT
living room.
He walks/limps/jogs/dances from the kitchen into the living room.
c. Er luft/hinkt/joggt/tanzt in das/*dem Zimmer von einem anderen Zimmer.
He walks/limps/jogs/dances in the.ACC/*the.DAT room from a.DAT different
room.
He walks/limps/jogs/dances into the room from a different room.
39
40
41
regarding whether being on the border is considered being inside or outside the landmark, and
thus there is no reason to reject the standard generalization based solely on this data.
5.4.2
Nested Landmarks
I now consider an unusual case where the figure in motion is located within a nested landmark
and motion occurs crossing the boundary of the innermost landmark, but not the outermost
landmark. The specific context of nested landmarks, as seen in (60), at first glance seems to
exhibit alternating features similar to the divergence attested in languages other than German
where both the accusative and dative are acceptable markings to describe the same event.
Though they describe the same event physically happening, (60a) emphasizes the border of the
forest (thus no border crossing) and (60b) emphasizes the border of the car (thus a border
crossing).
(60) a. CONTEXT : Car is on a road in forest.
Vom Auto, luft/spaziert/wandert/bummelt sie im Wald.
From the car, walks/strolls/wanders/ambles she in the.masc.DAT forest.
From the car she walks/strolls/wanders/ambles inL the forest.
b. CONTEXT : Car is on a road in forest.
Vom Auto, luft/spaziert/wandert/bummelt sie in den Wald.
From the car, walks/strolls/wanders/ambles she in the.masc.ACC forest.
From the car she walks/strolls/wanders/ambles into the forest.
When examining the semantic consequences for the alternate case use, it is clear that the choice
of accusative and dative reveal conflicting interpretations. The use of the dative in (60a) is not
referring to the forest as the goal of a motion, rather it is emphasizing that the movement of
exiting the car is located within the forest, compared to the emphasis of exiting the car in (60b).
Thus nested landmark events are not evidence of directional reinterpretations of goals.
42
Zwarts categorizes parken as a verb of directed motion (Zwarts 2006 : 5), but in comparison
with other verbs that encode directionality, it does not function similarly (in section 5.5.2 I argue
that it is in fact not a verb of directed motion). Verbs of directed motion, such as werfen to
throw or fallen to fall, entail a trajectory, but their interpretation as either directed or locative
can vary with proximity and spatial configuration, as is suggested in (61).
(61) Fallen to fall
CONTEXT : Man standing on the table, then falls onto the table
a. Der Mann fllt auf *den/dem Tisch.
The man falls on *the.ACC/the.DAT table.
The man falls onL the table.
CONTEXT : Man falls from a second floor balcony.
b. Der Mann fllt auf den/*dem Tisch.
The man falls on the.ACC/*the.DAT table.
The man falls onto the table.
While (61a) prefers a locative interpretation, it is truly a locative reading and not reinterpreted to
mean the goal of motion. When the source of directed motion is exclusively distinct from the
goal of motion, the sentence follows the standard generalization and only the accusative is
acceptable.18
18
There are four, well-documented directional-PP selecting verbs that can be considered a subset
of directed motion verbs: stellen, legen, setzen, and hngen. Not only are they verbs of directed
43
Zwarts claims that parken is also a directed motion verb, and therefore that it does not
require the use of accusative goal marking. He claims, in fact, that this marking is redundant,
which is why the goal constituents of parken use the dative marking. But parken does not behave
in the same way other verbs of directed motion (like fallen or werfen) behave above. Unaffected
by context, parken uniformly selects a location (dative) PP, as evidenced in (62).
(62) a. Er parkt das Auto vor *das/dem Haus.
He parks the car before *the.ACC/the.DAT house.
He parks the car in front ofL the house.
b. Er stellt das Auto vor das/*dem Haus.
He puts the car before the.ACC/*the.DAT house.
He parks the car in front ofG the house.
(German Zwarts 2006 : 6 (9b))
This same directionality pattern is evidenced in English, as seen in (63) and (64).
(63) a. He parked the car inL the garage
b. He parked the car *into/inG the garage.
(64) a. He put the car ?inL the garage.
b. He put the car into/inG the garage.
Discussion with a native speaker confirmed that the action described by parken was not merely
putting the gear into park (which would suggest a purely locational reading and thus restricted
dative marking), as there is some implied motion involved (see (65b)) and you can apply the
verb to non-car figures (see (65c)).
(65) a. CONTEXT : Man is stopped in the car in the driveway, puts gear into park
Er parkt das Auto vor *das/dem Haus.
He parks the car before *the.ACC/the.DAT house.
He parks the car in front ofL the house.
b. CONTEXT : Man drives into the driveway then parks
Er parkt das Auto vor *das/dem Haus.
He parks the car before *the.ACC/the.DAT house.
He parks the car in front ofL the house.
motion, but also they inherently require a goal constituent in the form of a directional
(accusative) PP.
44
46
Russian verb for to park). Israeli does not hypothesize as to the source of this unexpected
marking, but I explore it here in the context of German.
(69) a. *On povesil rubaku vo dvorA.
he.NOM hung.PG.TR.M shirt.ACC.SG in courtyard.ACC.SG
He hung his shirt in the yard.'
(Russian Israeli 2004 : 5 (7))
b. On povesil rubaku vo dvoreL.
He.NOM hung.PF.TR.M shirt.ACC.SG in courtyard.LOC.SG
He hung his shirt in the yard.
There are several ways to describe this particular event in German using a set of verbs
(aufhngen to hang up, suspend hngen to hang, legen to lay/set) and a set of goals (der Hof
the yard, die Leine the line, die Wand the wall). I have tested several iterations of these
verb-goal combinations in (70)-(72) to isolate the source of the difference.
(70) Aufhngen to hang/suspend sth.
a. Er hngt sein Hemd auf *den/dem Hof auf.
He hangs his shirt on *the.ACC/the.DAT yard up.SP
He hangs his shirt up inL the yard.
b. *Er hngt sein Hemd auf die/der Leine auf.
He hangs his shirt on the.ACC/the.DAT line up.SP.
He hangs his shirt up onL the line.
c. Er hngt das Bild an *die/der Wand auf.
He hangs the picture on *the.ACC/the.DAT wall up.SP.
He hangs the picture up onL the wall.
(71) Hngen to hang
a. Er hngt sein Hemd auf den/*dem Hof.
He hangs his shirt on the.ACC/*the.DAT yard.
He hangs his shirt inG the yard.
b. Er hngt sein Hemd auf die/*der Leine.
He hangs his shirt on the.ACC/*the.DAT line.
He hangs his shirt onG the line.
c. Er hngt das Bild an die/*der Wand.
He hangs the picture on the.ACC/*the.DAT wall.
He hangs the picture onto the wall.
47
48
change-of-state by motion verbs, but the goal of the motion that the verb is describing the end of
is encoded in the verb itself. Further evidence of this verb class is exhibited by (73).
(73) Ankommen to arrive
a. Er kommt an *die/der Station an.
He comes on *the.ACC/the.DAT station on.SP.
He arrives atL the station.
b. Er kommt unter *die/der Decke an.
He comes under *the.ACC/the.DAT canopy on.SP.
He arrives underL the canopy.
c. Von Frankreich, kommt der Mann an *den/dem Berliner Bahnhof an.
From France, comes the man on *the.ACC/the.DAT Berlin Station on.SP.
From France, the man arrives at the Berlin Station.
This may be more of a historical note on the development of the verb, but it has important
connections to the role of syntactic elements and post-positional particles to be discussed in
section 5.6.
5.5.3
Verb Summary
Subset
Directed-motion
Verbs
Dir.PP-selecting
Verbs
Change-of-state by
motion verbs
Goal-encoding Verbs
Verbs
Case
Fallen to fall,
werfen to throw
Accusative or Dative,
Depending on
Reading
Accusative
hngen to hang,
stellen to put, legen
to lay, setzen to sit
parken parken,
verstecken to hide,
verbergen to hide,
erscheinen to appear
aufhngen to hang
up, suspend,
ankommen to arrive
Dative
Dative
49
There are two main types of verbs I have explored, directed-motion verbs and change-of-state by
motion verbs, which are subsequently divided into four distinct categories. One set of directed
motion verbs select for a directional PP as a goal of motion, while another set of directed motion
verbs can select for a directional or locational PP depending on context. Of the change-of-state
by motion verbs, one set describes the location of the resulting state of motion through
lexicalization and the other set describes it through a compound construction. Crucially, though,
in free alternation dative never comes to encode a goal the interpretation is always locational,
and dative is judged unacceptable if this is not compatible with the context, rather than
undergoing reinterpretation as in other languages.
Directional Particles
Directional particles in German, like other languages mentioned in section 3.4, serve to explicate
the context of an event and are a cornerstone of directional context. Abraham (2003) even argued
that the German alternating prepositions did not actually alternate, that these prepositions only
governed the dative and the use of accusative and implied directionality is derived from either an
overt or covert (i.e. unexpressed, but semantically present) deictic particle. While this explication
of context in languages other than German by use of directional particles allows for a directional
reinterpretation of a location, the explicit directional context in German instead necessitates the
use of the accusative.
Nikitina (2008 : 181) notes that directionally ambiguous prepositions in English were
made explicitly directional with the insertion of a directional element, as in (74).
50
(74) If I dont come back in the house, Breeds going to (Nikitina 2008 : 181 (8b))
The addition of back serves to explicate the source of motion as exterior to the house, making the
source and goal different and in is reinterpreted as a directional PP. The examples in (75) show
the addition of zuruck back to similarly explicate the source of motion as distinct from the goal.
Unlike English however, the locative marking is disallowed rather than reinterpreted.
(75) Goal-marking particles (zuruck back)
a. Der Mann luft zuruck in das/*dem Zimmer.
The man walks back in the.ACC/the.DAT room.
The man walks back into the room.
b. Der Mann schwimmt zuruck unter das/*dem Wasser.
The man swims back under the.ACC/*the.DAT water.
The man swims back underG the water.
c. Es geht zuruck in das/*dem Wasser.
It goes back in the.ACC/*the.DAT water.
It goes back into the water.
d. Es geht zuruck unter das/*dem Wasser.
It goes back under the.ACC/*the.DAT water.
It goes back underG the water.
German has complementary locational particles that similarly specify the context of a motion,
shown in (76), and explicate the source of movement as the same as the goal of movement.
(76) Locational particles
a. Der Mann geht in *das/dem Zimmer auf und ab/hin und her/herum.
The man goes in *the.ACC/the.DAT room there and back/here and there.
The man goes back and forth (paced) inL the room.
Whether directional or locational in nature, German particles serve to explicate context where
not otherwise explicit and the consequent case is used according to the acceptable interpretations.
German post-positional particles are another subset of directional particles and must be
placed at the end of a sentence. The pairs of sentence presented in (77-9) illustrate how the post-
51
positional particles explicate the context, necessitating the use of accusative when the particle is
included.
(77) a. Sie luft in die/der Straenbahn.
She walks in the.ACC/the.DAT streetcar.
She walks into/inL the streetcar.
b. Sie luft in die/*der Straenbahn hinein.
She walks in the.ACC/*the.DAT streetcar into.
She walks into the streetcar.
(78) a. Sie klettert in die/den Bamen.
She climbs in the.ACC/the.DAT trees.
She climbs into/inL the trees.
b. Sie klettert in die/*den Bamen hinauf.
She climbs in the.ACC/*the.DAT trees up.
She climbs up into the trees.
(79) a. Sie klettert auf den/dem Berg.
She climbs on the.ACC/the.DAT mountain.
She climbs onto/onL the mountain.
b. Sie klettert auf den/*dem Berg hinauf.
She climbs on the.ACC/*the.DAT mountain up.
She climbs up onto the mountain.
The inclusion of a postpositional particle serves primarily to specify directionality where it is not
already clarified. The use of steigen in (80) disallows the dative, as opposed to klettern, due to
lexically encoded telicity. The type of movement described by steigen involves smaller climbing
motions like climbing a ladder or into a car that has a specified end. The inclusion of a postpositional particle in (80b) merely serves to emphasize the directionality of the parent sentence.
(80) a. Sie steigt auf den/*dem Berg.
She climbs on the.ACC/*the.DAT mountain.
She climbs onto the mountain.
b. Sie steigt auf den/*dem Berg hinauf.
She climbs on the.ACC/*the.DAT mountain up.
She climbs up onto the mountain.
The function of post-positional particles in German corroborates Nikitinas assertion that
directional particles specify directional contexts to allow exclusively directional readings.
52
However, this effect is limited to contexts like (77)-(79) where either interpretation is possible
given the use of the proper case. The inclusion of directional particles has no tangible effect on
contexts where directionality is already implied by presented contexts. Rather than allowing the
directional reinterpretation of the dative case, the inclusion of a directional particle necessitates
the use of the accusative. Still, the larger point is that when the reading is directional, accusative
becomes obligatory.
5.6.2
With the set of alternating prepositions, the complement case choice traditionally marks goal or
location PPs. However, there is evidence of case choice marking a third argument: the entire
path. While the interpretation of goals and locations in German through case choice has so far
proved unambiguous, the strategy for encoding a complete path through case choice does not
exhibit similar rigidity.
The following pattern is exhibited through the combination of case and post-positional
particles in (81): ACC marks the goal of motion as under the bridge, DAT marks the location of
motion as under the bridge, DAT+PostPos particle marks the path of motion as under the bridge,
and ACC+PostPos particle is unacceptable.
(81) a. Er schwamm unter die Brcke.
He swam under the.ACC Bridge.
He swam underG the bridge.
b. Er schwamm unter der Brcke.
He swam under the.DAT bridge.
He swam underL the bridge.
c. Er schwamm unter der Brcke hindurch.
He swam under the.DAT bridge through.
He swam through the under side of the bridge.
d. #Er schwamm unter die Brcke hindurch.
He swam under the.ACC bridge through.
He swam through the under side of the bridge.
53
54
While in (83a) the airspace above the city is indeed the only location of motion, the airspace is
not the goal of motion in (83b); it is the path. Where traditionally the accusative marks the goal
of motion, here it marks the path. The DAT+PostPos particle is used to mark a path in (81c)
while ACC+PostPos particle is used to mark a path in (82c) and (84).
Yet another alternation of path encoding is seen in (85). In this case, the path of walking
across the bridge can be conveyed by ACC (85a), ACC+PostPos particle (85c), or DAT+PostPos
particle (85d). (85c) emphasizes the bridge, while (85d) emphasizes the action of walking across.
(85) a. Er luft ber die Brcke.
He walks over the.ACC bridge.
He walks over the bridge and stops in the middle.
-ORHe walks over the bridge to the other side.
b. Er luft ber der Brcke.
He walks over the.DAT bridge.
He is walking in some space above the bridge (e.g. on a tightrope).
c. Er luft ber die Brcke hinber.
He walks over the.ACC bridge across.
He walks across the bridge.
d. Er luft ber der Brcke hinber.
He walks over the.DAT bridge across.
He walks across by way of the bridge.
Instead of exhibiting an alternation in the encoding of goals, German exhibits an alternation in
the expression of path. While it is not the sought after directional reinterpretation of a location,
German does exhibit an alternation in path encoding given the use of a dative PP (as in (81c) and
(85d)) and accusative PPs (as in (82c), (83b), (84), (85a), (85c)) being both interpreted as the
55
path of motion. It is clear that the encoding of path of motion through case in German deserves
further study.
56
6. CONCLUSION
Given explicit context or intention, German nonetheless still adheres to the standard
generalization of case choice with alternating prepositions. Through two methodologies, this
study strongly suggests that German does not employ the directional reinterpretation of locations
as goals exhibited across other typologically diverse languages. Given the breadth of languages
across typologies that employ some version of this pragmatic reinterpretation, the absence of this
strategy in German as determined by this study is striking and necessitates the reconsideration of
this pragmatic tool as a natural phenomenon of language, placing limits on the assertions by
Tham et al (2012) and Beavers et al (2010). Though this study did not find evidence for the
pragmatic reinterpretation of locations as goals, it did provide evidence that German has a richer
morphological system for encoding direction and path, whether through prepositions, case,
compound prefix verbs, or post-positional particles, than is typically discussed.
It is tempting to address the reason why, of all languages, German does not allow this
widely attested exception. One theory may point to the fact that many languages exhibiting this
alternation involve some amount of explicit additional information (in is shorter than into) where
the locational variant is shorter and easier to say, but in German the locational variant is not
actually unmarked in any obvious way, thus there is no pragmatic reason to use it over the
directional variant. Such an assertion would also conflict with the attested reinterpretations
available in Russian, which uses a case marking system almost identical to German. The
locational variants in Russian often include more information than the directional variants,
further conflicting with any lexically based pragmatic use.
57
The simplest and likeliest reason draws on parametric variation: the idea that there is a set
of tools languages can choose from to express ideas and German just opts out of this pragmatic
variation, similar to how it opts into a case system. But this motion specific parametric variation
is itself a novel finding in the study of motion typologies. A key tenant allowing for the
pragmatic reinterpretation evidence presented by Beavers et al (2010) centers on their
redefinition of Talmyan typologies. Specifically, they claim that the apparent categories of
motion encoding typologies are not guided by any motion specific parameters; rather motion
encoding is guided by a broader set of lexical, morphological, and syntactic grammatical devices
available to languages, which are exploited by languages in encoding motion and thus derive
apparent motion encoding typologies, defining these typologies as actually based on more
general grammatical resources. The findings from this study, however, suggest that there is at
least one motion specific parameter available to languages: the ability to reinterpret locations as
goals. Unless this parameter derives from some deeper interpretive process, which has yet to be
determined, the pragmatic reinterpretation of locations as goals is a motion specific parameter
available to languages when encoding motion.
The impact of this study is two-fold. Firstly, it provides evidence of the first known
exception to the proposed universal ability of languages to reinterpret locations as goals.
Secondly, it provides evidence that there is in fact at least one motion specific parameter that
guides motion-encoding strategy in language. The natural question following this study is
whether there are other languages showing a similar rigidity in goal encoding and whether there
are other motion specific parameters that guide motion-encoding strategies available to
languages. I hope this study of the interpretation of German directional prepositional phrases will
58
provide groundwork for other studies of motion-encoding strategy and motion specific
parametric variation.
59
7. BIBLIOGRAPHY
Abraham, Werner. 2003. The myth of doubly governing prepositions in German. In E. Shay &
U. Seibert (eds.), Motion, Direction and Location in Languages. In honor of Zygmunt
Frajzyngier. 19-38. Amsterdam, Philadelphia: John Benjamins. (Typological Studies in
Language 56).
Baroni,M., S. Bernardini, A. Ferraresi and E. Zanchetta. 2009. The WaCky Wide Web: A
Collection of Very Large Linguistically Processed Web-Crawled Corpora. Language
Resources and Evaluation 43 (3): 209-226.
Beavers, John, Beth Levin, Shiao Wei Tham. 2010. The Typology of Motion Expressions
Revisited. J. Linguistics 46: 331-377.
Belichova-Krzhizhkova, E. 1974. Prostranstvennaia determinaciia v russkom i v cheshskom
iazykax. M. Zatovkaniuk, T.I. Konstantinova and A.G. Shirokova (eds.) Konfrontan
stadium rusk a esk gramatiky a slovn zsoby: Sopostavitelnoe izuchenie grammatiki
i leksiki russkogo i cheshskogo iazykov, vol. 1. 92-118. Prague: Universita Karlova.
Cienki, Alan J. 1989. Spatial Cognition and the Semantics of Prepositions in English, Polish,
and Russian. Munich: Verlag Otto Sagner.
Dikken, Marcel den. 2003. On the syntax of locative and directional adpositional phrases. Ms.
The Graduate Center of the City University of New York.
Estigarribia, Bruno & Beth Levin. 2007. When can UNDER mean *UNDER TO? Eighth Annual
Semantics Fest, Stanford University: March 2007.
Fbregas, Antonio. 2007. The exhaustive lexicalization principle. Nordlyd: Troms Working
Papers in Linguistics 24: 165-199.
Gehrke, Berit. 2006. On directional readings of locative prepositions. To appear in the
proceedings of ConSOLE XIV in Vitoria-Gasteiz, 2005.
Haider, Hubert. 2010. The Syntax of German. New York: Cambridge University Press.
Israeli, Alina. 2004. Case choice in placement verbs in Russian. Clossos. Issue 5 (Summer
2004): 1-54.
Langacker, Ronald W. 1987. Foundations of Cognitive Grammar, vol. 1. Stanford: Stanford
University Press.
Nedashkivska, Alla. 2001. Wither or where: Case choice and verbs of placement in
contemporary Ukrainian. Journal of Slavic Linguistics 9: 213-51.
60
Nikitina, Tatiana. 2007. Variation in the encoding of goals and static locations in Russian. Ms.
Stanford University.
Nikitina, Tatiana. 2008. Pragmatic factors and variation in the expression of spatial goals. In A.
Asbury, J. Dotlacil, B. Gehrke, R. Nouwen (eds.), Syntax and Semantics of Spatial P,
175-195. Amsterdam/Philadelphia: John Benjamins Publishing Company.
Riemsdijk, Henk van. 2007. Case in Spatial Adpositional Phrases: The Dative-Accusative
Alternation in German. In G. Alboiu, A. Avram, L. Avram, & D. Isac (eds.) Festschrift
for Alexandra Cornilescu,
Talmy, Leonard. 1985. Lexicalization patterns: semantic structure in lexical forms. In T. Shopen
(ed.), Language Typology and Syntactic Description, vol. 3: Grammatical Categories
and the Lexicon, 57-149. Cambridge: Cambridge University Press.
Talmy, Leonard. 2000. A Typology of Event Integration. Towards a Cognitive Semantics:
Typology and Processing Concept Structuring. Cambridge, MA: MIT Press.
Tham, Shiao We, Beth Levin, John Beavers. 2012. Directional interpretations with locative
prepositions. Presented at The Meaning of P, Bochum, Germany: November 2012.
Thomas, E. 2001. On the expression of directional movement in English. Essex Graduate Papers
in Language and Linguistics 4: 87-104.
Thomas, E. 2005. On syntactic versus semantic telicity: Evidence from In and On. In H.
Cuyckens, W. de Mulder & T. Mortelmans (eds.) Adpositions of Movement, 145-166.
Amsterdam: John Benjamins.
Ungermanov, Marta. 2005. Locative complements of verbs of movement in Czech: Some
typical structures and their interpretation. H. Cuyckens, W. de Mulder, & T. Mortelmans
(eds.) Adpositions of Movement. 87-113. Amsterdam: John Benjamins.
Zaitseva, Valentina. 1994. The metaphoric nature of coding: Toward a theory of utterance.
Journal of Pragmatics 22. 103-126.
Zwarts, Joost. 2006. Case Marking Direction: The Accusative in German PPs. Presented at the
Variation and Stability in Grammar workshop. Nijmegen: Radaboud University
Nijmegen, April 2006.
Zwarts, Joost. 2010. A hierarchy of locations: evidence from the encoding of direction in
adpositions and cases. Linguistics 48: 983-1009.
61
8. BIOGRAPHY
Amanda Haight was born on June 2, 1993 and raised in Austin, Texas. She attended the Liberal
Arts and Science Academy of Austin from 2007-2011. Upon graduating, she enrolled at the
University of Texas at Austin with majors in Plan II Honors and Linguistics. During her time at
UT, she served as president and co-founder of the Undergraduate Linguistics Society, an
Associate Editor for The Daily Texan, participated as an actor in the UT Theatre Department,
and was an active member of Texas Spirits. Amanda graduated Phi Beta Kappa and Phi Kappa
Phi in the fall of 2015 and will join Bain & Company as an Associate Consultant in Fall 2016.
62