Monique Lambert

1
Empirical approaches to aspectLanguage systems
and learner languages

• Monique Lambert
• Christiane von Stutterheim
• Mary Carroll
• Monique Flecken
• University of Heidelberg
• Université Paris VIII

ICLC 6th FU Berlin, 30.9. 2.10.2010
2

• Your analysis of the English present perfect is
  by far the best. You just don't use it properly.
• Personal communication between Dan Slobin and
  Wolfgang Klein

3
The problem

• Aspectual categories are extremely hard to
  acquire in a second language. Even at very
  advanced stages we find on the side of the
  learner
• - inconsistent usage
• - absence of native-like intuitions
• Aspect hypothesis for L2 acquisition
• Progressive aspect is related to inherent
  atelicity
• on the side of the native speaker
• - no explicit knowledge
• What exactly is this knowledge?

4
The progressive a point in question

• Context what is Peter doing in the kitchen?.
• Peter is baking a cake
• Peter ist dabei einen Kuchen zu machen.
• Piet is een coekje aan het maken.
• Pierre est en train de faire un gâteau.
• Pietro sta facendo una torta
• Context  Can you see Peter over there ? Where is
  he going?
• Peter is walking to the station.
• ?Peter ist dabei zum Bahnhof zu gehen.
  (inchoative reading?)
• ?Piet is naar de station aan het lopen.
• Pierre est en train daller à la gare.
• Pietro sta andando a la stazione.

5
How do we know?
The multilingual research group
Abassia Bouhaous Natascha Sahonenko Barbara
Schmiedtová Takara Baumbach Xingyu Zhu
Ginés Carpena Renate Delucchi Naoko Tomita Silvia
Natale

• Bergljot Behrends
• (University of Oslo)

6
The state of research

• We have legions of studies on temporal
  categories and we still cannot answer basic
  questions when learners want to understand and
  manage e.g. the English or Japanese aspectual
  system. If the semantic descriptions of
  tense/aspect-forms were as straightforward as
  suggested by many theoretical accounts then this
  should provide the basis for using these forms
  appropriately.

7
The question

• In acquiring a second language we not only
  acquire new forms and new functions. We also have
  to acquire knowledge as to how grammatical
  categories are put to use.
• In the case of aspect what are the principles
  that a native speaker relies on when selecting a
  particular temporal perspective?
• Very few studies have addressed this question
  crosslinguistically.

8
Question cont.

• Theoretical accounts do not explain different
  constraints across languages in the use of one
  semantic operation (e.g progressiveness). They do
  not explain differences within one verb type
  (e.g. Vendler accomplishment / achievement) as
  being subjected to one aspectual operation.
• They do not explain principles underlying
  optional marking.

9
Goals of the comparative study

• How can we describe the differences between
  languages within what is described as one
  aspectual category in current theories?
• How can we theoretically account for developing
  systems and the optionality of aspectual marking
  at a given stage?
• How can we describe the principles of use of
  aspectual forms?
• How and to what extent do second language
  learners discover structure and function of
  aspectual systems?

10
Empirical approach to language use the basic
idea

• Speakers see the same real world scene, they are
  introduced to the verbal task under the same
  temporal perspective.
• This allows for a systematic variation of event
  features potentially relevant for temporal
  perspective taking and selection of aspectual
  marking.
• (Elicitation instruction Say what is
  happening/what happens/happened has no impact.)

11
The levels of event construal
External world the stimulus
selective attention
perceptual filter
Conceptual representation
information selection, perspective taking
cognitive filter
Verbal representation
12
Scenes
13
Scenes
14
Scenes
15
Design of the studyvariables language,
situation type

• Languages investigated
• Adult speakers L1 (similar education, aged
  between 20 - 30) of
• Germanic (German, Dutch, Norwegian, English),
• Romance (French, Italian, Spanish),
• Slavic (Russian, Czech, Polish, Bulgarian),
• Arabic (Modern Standard, Algerian Arabic),
  Japanese, Chinese

Adult speakers L2 (matched for sociocultural
background), (very) advanced
L2 German English/Japanese/Italian French
L1 English/Russian/Czech German Dutch
16
Design of the studyvariables language,
situation type

• Situation types investigated
• Motion events
• with endpoint reached (somebody is walking to a
  house)
• with endpoint not reached (somebody is walking on
  a road)
• Causative action
• (somebody is knitting a scarf)

16
17
Type of Data

• Language production audio-taped, transcribed
• (Speech onset time
• Eye tracking
• Memory performance)

18
Range of languages

• Differences of aspectual systems with respect to
  the degree of grammaticalisation

19
Hypothesis Attractor factors for progressive
aspect

• type of situation
• Locomotion (spatial orientation) walk, causative
  action (with affected / effected object) paint a
  picture vs. build a model airplane, activity
  (swim, surf)
• relevant features of situations
• /- endpoint of situation bake a cake vs. walk
  on a lawn
• /- endpoint reached a man walks into a house vs.
  a man walks towards a house
• /- homogeneity knitting a scarf vs. tidying up a
  room
• Material 40 /80 videoclips, critical items,
  control items, presented in randomised order

20
Results fully grammaticalised vs. lexical
Use of aspectual on-goingness marker in the given
task across all situation types (present tense)
21
Developing systems the role of transition points

• Situations with a qualified resultant state
• (building a monument moulding a vase painting a
  picture knitting a scarf)
• (20 speakers per language 10 situations)

22
Developing systems the role of transition points

• Situations without an inherent boundary
• surfing, jogging, kicking a football, swimming
• (20 speakers 10 situations)

23
Developing systems the role of transition points

• motion events figure underway endpoint not
  reached
• (20 speakers 10 situations)

24
Situation and aspect general tendency

• A progressive perspective is more likely to be
  selected in situations with an inherent
  transition point and a qualified resultant state,
  compared to those which do not show a temporal
  property of this kind.

25
Developing systems the role of homogeneity

• Situations with a qualified resultant state
• and homogeneous subevents
• (moulding a vase knitting a scarf)
• (20 speakers per language 5 situations)

26
Situation and aspectgeneral tendency

• The main attractor factors for progressive
  aspect are situations with a qualified second
  state and a first state with a dynamic action and
  homogeneous subevents.
• The relevance of this type of situation for view
  point selection may lie in the fact that it
  delivers a measure for progression toward a
  qualified resultant state that can be expected
  to come about (the moulded vase, for example).

27
Summary of empirical findings

• Crosslinguistic patterns are rooted in a limited
  set of temporal categories.
• Critical for PROG/IMPERF is the contrast of
  states and with this some point in time
  (boundary, closure, transition point homogenous,
  dynamic, durative, first state).
• Major attractor/emerging systems 2-state
  situations with change of quality resulting in a
  tangible second state
• Locomotion implies specific constraints across
  languages.

28
Conclusion Systems

• The empirical result show the relevance of
  preferences in language use. The selection among
  options in perspective taking is not arbitrary
  but guided by deeply rooted principles mediated
  through grammar. These draw on universal notional
  categories (such as quality of entities,
  delimitation, temporal phase) to form language
  specific clusters of related temporal features.
• These clusters are conceptually dominant and
  provide a language specific basis for temporal
  perspective taking
• ? principles of use

28
29
Implications for learning

• Aspect requires perspective taking.
• Languages differ with respect to the options at
  a fine grained level.
• At text level these options are constrained by
  principles of information organisation.
• ? Learning problems

30
The acquisition of aspect in L2

• Learning task
• forms
• functions
• principles of use, i.e. which context
  licenses/requires a particular grammatical
  operation
• Previous research
• Studies in SSLA focus on earlier phases of the
  acquisition of aspect and form-function
  relations.

31
Empirical study very advanced L2 speakers

• Same material as for the L1
• Generally similar results as the L1 speakers for
• German gt English
• Italian gt French
• German gt Japanese

32
Empirical study very advanced L2 speakers

• Problems arise
• at text level (language specific principles of
  perspective taking cf. Tomita)
• when speakers are distracted by
• time pressure
• inappropriate perspective introduction

33
L1 German/French L2 English text production

• Use of progressive aspect in film renarrations
• L1 English
• and he hears the water again
• but suddenly hes pushed up into the air by one
  of these towers (completion)
• that comes up out of the ground
• and hes up on top of this tower (state
  it is now the case)
• and hes looking around for the water
  (progressive ongoing event)
• L1 German ? English
• so he ends up at the summit of the rocks
• and is just trying to look around
• to see what's happening

34
L1 German/French L2 English text production

• combination of phasal segmentation with holistic
  perspectives
• L1 German and falls down
• like always he is never hurt and then he
  walks up to the water
• and the water is dripping on a huge stone
  plate
• L1 French he starts digging a hole eventually
  falls into it
• and is covered with the sand
• he is trying to get out of it
• but he cant
• and he falls

35
L1 German/French L2 English text production

• lack of anaphoric situation for events presented
  under an ongoing perspective
• L1 German there is no water in him
• he consists of sand
• then he is still searching for water
• L1 French and he starts digging
• and the ground is suddenly opening
• he suddenly watches little holes in the
  ground

36
Conclusion L2 acquisition

• In order to decide when the one or the other
  perspective has to be selected, world knowledge
  has to be combined with language concepts (like
  in the case of spatial prepositions).
• For the L2 learner there are no formal
  indicators which help to discover the principles
  of use.
• Under processing aspects L2 syntactic knowledge
  is less automated.

37
Bibliography

• Binnick, Robert I. (2006). Aspect and
  Aspectuality. In Bas Aarts April McMahon
  (Eds.), Handbook of English Linguistics. Oxford
  Blackwell. S.244-68.
• Klein, W. (1992). The present perfect puzzle,
  Language 68, 525-552.
• Klein, Wolfgang (1994). Time in Language. London
  Routledge.
• Tomita, Naoko (2008). Der Informationsaufbau in
  Erzählungen. (Japanisch, Deutsch, Englisch).
  München Iudicium.

37
38
Time argument structures (W. Klein)

• Temporal structure of a 1-state situation
• The situation in the external world is a person
  who sleeps
•  
• Entity 1 person e1
• ------------------------------------------------
  ----------------------
• prestate course of activity post state not
  specified
• no spec. homogeneous
• for e1 subintervals
• Temporal structure of a 2-state situation with an
  unqualified second state
• The situation in the external world is a person
  who walks up to a house
• Entity 1 person e1, entity 2 house/goal e2
•  
• ------------------------------------------------
  ----------------------
• prestate course of activity goal reached, end
  of activity
• no spec. homogeneous no resultant state
  qualified
• for e1, e2 subintervals only e1 at loc e2
• e1 active

39
Time argument structures (W. Klein)

• Temporal structure of a 2-state situation with
  effected object
• The situation in the external world is a person
  who bakes a cake
• Entity 1 person e1, entity 2 cake e2
•  
• -----------------------------------------------
  -----------------------
• prestate course of activity resultant state
• no spec. not homogeneous
• for e1, e2 subintervals
• e1 active e2 exits
• e2 emerges no spec. for e1
•