Limiting writing task freedom by constraining Cmap link type

1
Limiting writing task freedom by constraining
Cmap link type

• Lawrie Hunter
• Kochi University of Technology
• http//www.core.kochi-tech.ac.jp/hunter/

2
BACKGROUND 2 PASSIONS
HAWW How Academic Writing Works (Information
strategy and readability)
MAPPING Foundation Information Structures (Knowled
ge Structures)
3
BACKGROUND 2 PASSIONS
1. HAWW
2. MAPPING
4
BACKGROUND 2 PASSIONS
1. HAWW
2. MAPPING
5
BACKGROUND 2 PASSIONS
1. HAWW
2. MAPPING
6
FOREGROUND 'Tensions'
Typical tensions in CALL work Learner
PC Learner software Learner target
content Learner interface Instructor
intervention learner performance Content
presentation style learner performance Ubiquity
learner motivation
7
FOREGROUND 'Tensions'
Hunter's tensions of interest Interface/task
learner perception of curriculum Representation
message comprehension Processing type learner
persistence Processing variation learning
effectiveness/efficiency Use of metalanguage
learner attack style Representation type
cognitive load in task scenario Representation
type degree of abstraction curriculum
transparency Representation type degree of
abstraction task success
8
FOREGROUND 'Tensions'
Hunter's tensions of interest Interface/task
learner perception of curriculum Representation
message comprehension Processing type learner
persistence Processing variation learning
effectiveness/efficiency Use of metalanguage
learner attack style Representation type
cognitive load in task scenario Representation
type degree of abstraction curriculum
transparency Representation type degree of
abstraction task success
9
FOREGROUND 'Tensions'
Global tension in Hunters mapping work   When
is mapping more effective/efficient than
text in task presentation/performance?  
10
FOREGROUND 'Tensions'

• Uses of Concept Mapping (really)
•  
• In scenarios where text is
• too dense or too clumsy
• for easy learner access to information
  structures.
•  
• 2. For abstraction.
•  
• 3. For a focus on information-linking functions.
•  
• 4. To downplay text orchestration.

11
Found application Constructivist L2 learning
task
12
Found application Constructivist L2 learning
task
Open-ended and expressive tasksfree the
learner to use language that is immediately,
personally relevant.
13
Found application Constructivist L2 learning
task
Open-ended and expressive tasksfree the
learner to use language that is immediately,
personally relevant.
My favorite game. A day in my life. How I
would reduce Japans CO2 emissions.
14
Found application Constructivist L2 learning
task
Open-ended and expressive tasksfree the
learner to use language that is immediately,
personally relevant.
BUT they also allow the learner to write freely
in terms of -content -information structure
and -rhetorical structure.
15
Found application Constructivist L2 learning
task
Open-ended and expressive tasksfree the
learner to use language that is immediately,
personally relevant.
BUT they also allow the learner to write freely
in terms of -content -information structure
and -rhetorical structure.
Not alwaysa good thing
16
How can we present taskso as to force certain
language behaviors?
e.g. use of certain structures e.g. use of
certain types of cohesion device e.g. use of
certain register
17
How can we present taskso as to force certain
language behaviors?Information mapping
systemsallow us to usefully present task
content in a low-text manner.
e.g. use of certain structures e.g. use of
certain types of cohesion device e.g. use of
certain register
18
Many information mapping systems

• Mind mapping
• Directed link maps
• Textured-link maps
• Horns argument mapping
• AusThinks argument mapping
• Rationale argument mapping
• RST maps
• Rhetorical Structure Theory

19
Quantum levels of mapping
Argument mapping
Information structure mapping
Syntactic mapping
Grammatical mapping (pseudo)
Association mapping
20
Mind mapping á la Tony Buzan
21
Mind mapping

• The links are all associations
• -i.e. zero granularity

22
Directed-link maps
http//www.inspiration.com/
23
Textured-link maps
Make steam
Rotate turbines
Generate electricity
Boil a liquid
seawater heat
fossil or N-heat
boil NH3
boil H2O
!
OTEC plants
older type plants
steam 20C
steam 500C
!
low power
high power
!
zero energy cost
high energy cost
graphically textured
24
Textured-link maps
textually textured
25
Textured-link maps (directed links)
v
26
Novakian maps
The basic Novakian concept map... usually
starts with a general concept at the top of the
map, and then works its way down ... to more
specific concepts.
Abrams, R. An Overview of Concept Mapping. In
Meaningful Learning A Collaborative Literature
Review of Concept Mapping. Retrieved March 18,
2008 at http//www2.ucsc.edu/mlrg/clr-conceptmappi
ng.html
27
When is a map Novakian?
The basic Novakian concept map... usually
starts with a general concept at the top of the
map, and then works its way down ... to more
specific concepts.
Concepts are placed in boxes...
Abrams, R. An Overview of Concept Mapping. In
Meaningful Learning A Collaborative Literature
Review of Concept Mapping. Retrieved March 18,
2008 at http//www2.ucsc.edu/mlrg/clr-conceptmappi
ng.html
28
When is a map Novakian?
The basic Novakian concept map... usually
starts with a general concept at the top of the
map, and then works its way down ... to more
specific concepts.
Concepts are placed in boxes... Lines are
drawn from a concept to a linking word to a
concept.
Abrams, R. An Overview of Concept Mapping. In
Meaningful Learning A Collaborative Literature
Review of Concept Mapping. Retrieved March 18,
2008 at http//www2.ucsc.edu/mlrg/clr-conceptmappi
ng.html
29
When is a map Novakian?
The basic Novakian concept map... usually
starts with a general concept at the top of the
map, and then works its way down ... to more
specific concepts.
Concepts are placed in boxes... Lines are
drawn from a concept to a linking word to a
concept. Sequences of concepts and linking words
do not always form grammatically correct
sentences.
Abrams, R. An Overview of Concept Mapping. In
Meaningful Learning A Collaborative Literature
Review of Concept Mapping. Retrieved March 18,
2008 at http//www2.ucsc.edu/mlrg/clr-conceptmappi
ng.html
30
Novakian maps
http//cmap.ihmc.us/
31
Many information mapping systems

• Mind mapping
• Directed link maps
• Textured link maps
• Horns argument mapping
• AusThinks argument mapping
• Rationale argument mapping
• RST maps
• Rhetorical Structure Theory

• Mind mapping
• Directed link maps
• Textured link maps
• Horns argument mapping
• AusThinks argument mapping
• Rationale argument mapping
• RST maps

32
Novakian maps (Novak Cañas, 2006) are type 3
maps, having textured link dyadic elements.

• Associational
• Directed link
• Textured, somehow directed link

33
Novakian maps, with their textured link dyadic
elements, can be used to represent arguments.
The rules of operation for Novakian maps may
be defined as Rectangles (nodes) contain
concepts. Rectangles are joined by labeled lines
(propositions). Propositions are usually
uni-directional. The label in the middle of a
proposition is called a linking phrase.
34
Clarify this studys purpose in terms of content
Aim of this design process -to create tasks and
devices that aid the TAW learner in
perceiving/using information structures
-argument structure, -knowledge structure
(following Mohan (1986) and Hunter (2002)) and
-message type (central message, background
information).
35
Clarify this studys purpose in terms of form of
language/information
36
Novakian maps are suitable for low-text
representations of rhetorical moves in logical
argument. The rules of operation for Novakian
maps may be defined as Rectangles (nodes)
contain nouns/noun phrases. Rectangles are
joined by labeled lines (rhetorical device
signals). Propositions are usually
uni-directional. The label in the middle of a
proposition is called a linking phrase.
37
Argument mappingas an inroad to task
constraint-an extreme case ofsummarizing-pers
uasive rhetoric filtered out
38
Important levels of mapping
Argument mapping Information structure mapping
39
Horns argument mapping
http//www.stanford.edu/rhorn/index.html http//w
ww.macrovu.com/
40
AusThink argument mapping
http//www.austhink.com/
41
Rationale argument mapping
http//www.austhink.com/
42

• RST mapping

Bill Manns Rhetorical Structure Theory (RST)
uses various sorts of "building blocks" to
describe texts. The principal block type deals
with "nuclearity" and "relations" (often called
coherence relations in the linguistic
literature.)
www.sil.org/mannb/rst/
RST links are rhetorical devices.
43
Cmap tools
http//cmap.ihmc.us/
Novakian links are syntactic devices.
44
Study task constraint work Constraining learner
behavior
45
Study task constraint work Using link labels to
constrain learner behavior
Allow only links which signal the information
structures which constitute the register
Here the register is TAW, and the degree of
abstraction is maximum, i.e. argument.
46
Study task constraint work Using link labels to
constrain learner behavior
Allow only links which signal the information
structures which constitute the register
47
Study task constraint work
Phase 1 unconstrained associative
mapping Learners made association type concept
maps of the content of a text. To force
summarization, a limit of 10 nodes was imposed.
Performance was strongly uniform in terms of
node content. Generally the learners used the
lexical units (e.g. subject-verb) found in the
text. ppt 1 associational mapping
48
Study task constraint work
Phase 2 unconstrained Novakian mapping Learners
map the same text again, using Cmap
software. Object "try to represent the argument
in the Niiho paper," which was described in
the article. Maps were highly varied (chaotic
variation) in content and in link type.
49
Study task constraint work
Phase 3 constrained Novakian mapping Learners
mapped the text a third time, again using Cmap
software, but with only the argumentation links
shown below.
50
Study task constraint work
Phase 3 constrained Novakian mapping Learners
mapped the text a third time, again using Cmap
software, but with only prescribed argumentation
links. Learner product was quite uniform in
choice of links from the array above. Map
structure and content varied relatively little.
ppt 2 argument mapping
51
Study task constraint work
Phase 4 text-based argument analysis
(follow-up) Different text Same argument
structure Same link constraints Subjects 9 Sen
tences 3.55/subject Sentences with non-list
labels 0.44/subject Misuses of linking
phrases 0.77/subject
52
Study task constraint work
Mapping vs. text Mapping Subjects 13 Dyads
6.9/subject Dyads with non-list labels
0.8/subject Misuses of linking
phrases 2.85/subject Text Subjects 9 Sentence
s 3.55/subject Sentences with non-list labels
0.44/subject Misuses of linking
phrases 0.77/subject
53
Study task constraint work
Conclusion Using this label-constrained mapping
approach, the subjects did successfully separate
persuasive from information-bearing text. Of
course the mapping approach is not essential to
the acquisition of that skill.
54
Study task constraint work

• Conclusion
• The value of the mapping approach in argument
  analysis
• visual accessibility of the representation of
  text structure
• forced articulation of relations between argument
  elements
• greater speed (lower cog load?) than in text
  approach
• Informal observation of learner behavior in
  constrained-link scenarios suggests that
  constraining link content can lead TAW learners
  to accurate, minimal summarization of the
  arguments in TAW text WITHOUT INSTRUCTION.

55
Study task constraint work

• Questions for future study
• Constrained-link maps afford valuable support to
  TAW learner analysis of text.
• Do they also afford valuable support to TAW
  learner composition processes?
• 2. What is the optimal size and content of the
  array of linking phrases for the representation
  of TAW argumentation?
• 3. Does the ad hoc representation technique
  developed here have implications for a
  multi-level model of the text-based
  representation of information structures?

56
Future work
Embed Hunters info structure mapping in Novakian
argument mapping
57
Hunters info structure maps
Classification
Description
Degree comparison
Attribute comparison
Contrast
!
Sequence
Cause-effect
58
Power generating systems
General process
Make steam
Rotate turbines
Generate electricity
Boil a liquid
seawater heat
fossil or N-heat
boil NH3
boil H2O
!
OTEC plants
older type plants
steam 20C
steam 500C
!
low power
high power
!
zero energy cost
high energy cost
hunter systems
59
Hunters info structure maps
60
Future work
Embed Hunters info structure mapping in Novakian
argument mapping -and see what happens!
61
Thank you for your attention.
You can download this .ppt from http//www.core.ko
chi-tech.ac.jp/hunter/presns/constraint/
Lawrie Hunter Kochi University of
Technology http//www.core.kochi-tech.ac.jp/hunter
/
62

• Limiting writing task freedom by constraining
  Cmap link type
• In constructivist learning environments,
  particularly in relatively teacher-remote
  situations such as those involving Learning
  Management Systems (LMS), learner freedom can be
  an obstacle to the achievement of specific
  educational objectives. Similarly, the use of
  concept map systems based on associational links
  can result in excessive task performance freedom.
  The Cmap approach lthttp//cmap.coginst.uwf.edu/gt
  involves the use of labelled links and thus
  constrains learner behavior towards more
  articulate expression, but still the degree of
  task performance freedom is high. Only by
  constraining the nature of link labels can the
  materials designer push the learner to very
  specific language behaviors.
• A case in point in work with the informal
  technical academic English register (used for
  example in science magazines and presentation
  scripts), an important second language
  reader/writer skill is the ability to distinguish
  between (1) technical content and scientific
  argument and (2) motivating or persuasive
  rhetoric.
• This paper first lays out the literature in
  outline form, showing the range of information
  mapping and concept mapping techniques and their
  strengths and weaknesses in terms of educational
  document design.
• The central content of the paper is a
  demonstration of the author's adaptation of Cmaps
  and Cmap tools for L2 learner analysis of
  technical text and the subsequent writing of
  technical text using Cmaps for argument mapping.
  This approach resembles argument mapping as
  described at http//www.austhink.org/critical/page
  s/argument_mapping.html and rhetorical structure
  analysis, as defined at http//www.sfu.ca/rst/02an
  alyses/index.html
• but differs in that in this case link label
  content is intentionally constrained the
  resulting reduced writing freedom can be employed
  to force learners to write in very narrow
  sub-genres of technical academic writing. This
  not only constrains the learner's writing it
  also forces the task designer and the instructor
  to work in narrow niche sub-genres, which are
  arguably enabling in their specificity and
  clarity.
• 294 words

63
In constructivist learning environments,
particularly in relatively teacher-remote
situations such as those involving Learning
Management Systems (LMS), learner freedom can be
an obstacle to the achievement of specific
educational objectives. This presentation
examines information mapping and concept mapping
techniques and demonstrates the author's
adaptation of Cmap tools lthttp//cmap.coginst.uwf.
edu/gt for L2 learner analysis and writing of
technical text. Link label content is
intentionally constrained to force learners to
write in very narrow sub-genres of technical
academic writing. This also forces the task
designer and the instructor to work in narrow
niche sub-genres, which are arguably enabling in
their specificity and clarity.