M2D: Monolog to Dialog Generation for Conversational Story Telling

Abstract

Storytelling serves many different social functions, e.g. stories are used to persuade, share troubles, establish shared values, learn social behaviors, and entertain. Moreover, stories are often told conversationally through dialog, and previous work suggests that information provided dialogically is more engaging than when provided in monolog. In this paper, we present algorithms for converting a deep representation of a story into a dialogic storytelling, that can vary aspects of the telling, including the personality of the storytellers. We conduct several experiments to test whether dialogic storytellings are more engaging, and whether automatically generated variants in linguistic form that correspond to personality differences can be recognized in an extended storytelling dialog.

Full text

M2D: Monolog to Dialog Generation for
Conversational Story Telling
Kevin K. Bowden, Grace I. Lin, Lena I. Reed,
Jean E. Fox Tree, and Marilyn A. Walker
Natural Language and Dialog Systems Lab
University of California, Santa Cruz
{kkbowden,glin5,lireed,foxtree,mawwalker}@ucsc.edu
Abstract. Storytelling serves many different social functions, e.g. sto-
ries are used to persuade, share troubles, establish shared values, learn
social behaviors, and entertain. Moreover, stories are often told conver-
sationally through dialog, and previous work suggests that information
provided dialogically is more engaging than when provided in monolog.
In this paper, we present algorithms for converting a deep representa-
tion of a story into a dialogic storytelling, that can vary aspects of the
telling, including the personality of the storytellers. We conduct several
experiments to test whether dialogic storytellings are more engaging,
and whether automatically generated variants in linguistic form that
correspond to personality differences can be recognized in an extended
storytelling dialog.
Keywords: Analysis and Evaluation of Systems, ICIDS, Dialog, Natu-
ral Language Generation, Personality, Conversational Storytelling
1 Introduction
Storytelling serves many different social functions, e.g. stories are used to per-
suade, share troubles, establish shared values, learn social behaviors, and enter-
tain [33, 24]. Moreover, stories are often told conversationally through dialog [39,
38] where the telling of a story is shaped by the personality of both the teller
and the listener. For example, extraverted friends actively engage one another in
constructing the action of the story by peppering the storyteller with questions,
and by asking the listener to guess what happened [39, 38]. Thus the same story
can be told in many different ways, often achieving different effects [22,40].
A system capable of telling a story and then retelling it in different settings
to different audiences requires two components: 1) a deep representation of the
story and 2) algorithms that render the story content as different discourse in-
stantiations. A deep representation of the story’s content, often called the story
or fabula, must specifies the events, characters, and props of the story, as well
as relations among them, including reactions of characters to story events. This
is accomplished through est [31], a framework that bridges the story annotation
tool scheherazade and a natural language generator (NLG).

2 Authors Suppressed Due to Excessive Length
The discourse representation is the surface rendering of the fabula, an
instantiated expressive telling of a story as a stream of words, gestures or ex-
pressions [3, 8, 28, 29]. This paper presents m2d, a framework with algorithms
that manipulate the story content to retell the story as a conversational dialog
between two people. An example of the original, monologic, and dialogic telling
of the Garden Story is shown in Figure 1. Note that highlighted areas indicate
the same information being presented differently at different stages.
Garden Story: Original
Today when I arrived at my community garden plot, it actually looked like a garden. Not a weedy
mess with maybe some stuff growing in it if you know where to look. We had hit the typical mid-
summer mess of fast-growing weeds and no time to do anything about it. Plus all the rain had
made a huge swamp and it was hard to get a moment to work when it wasn’t actively pouring. I put
in a bunch of time this past week, and it’s paying off. Along with free-standing non-weed-choked
plants, I have now re-planted three of the beds with salad greens, spinach, and chard. And while
the viability of the seeds was questionable, I accidentally unearthed some from the bed I planted
2 days ago and they had already started to sprout! ... etc.
Garden Story: Monolog generated by est Garden Story: Dialog generated by m2d
The radishes charmed the butterflies. The
communal garden was weedy. It rained. The
communal garden was swampy. It rained. The
productive gardener planted the plants. The
gardener planted the chards the lettuces and
the spinach. The pleased gardener did not
expect for the chards the lettuces and the
spinach to grow. The chards the lettuces and
the spinach sprouted. The gardener mistakenly
dug the chards the lettuces and the spinach.
The surprised gardener saw for the chards the
lettuces and the spinach to sprout. The
communal garden was not weedy. The
communal garden was not swampy. The
communal garden was productive. The
gardener was proud. The eager gardener
wanted to reap the lettuces. The radishes were
droopy. The gardener planned to remove the
radishes. The thoughtful gardener thought the
flowers charmed the butterflies.
S1: The radishes charmed the butterflies. Techni-
cally, the damn communal garden was not weed-
less. Err ... it rained. Great right, it was really
swampy.
S2: Yeah, it rained. The productive gardener
planted the plants. I mean, she planted the chards
the lettuces and the spinach.
S1: She did not expect for them to grow. The
chards the lettuces and the spinach sprouted,
didn’t they?
S2: Unfortunately, the gardener mistakenly dug
them. She saw for the chards the lettuces and the
spinach to sprout.
S1: The communal garden was not weedy. It was
pretty productive and not swampy.
S2: Mmhm ... the gardener was proud and wanted
to reap the lettuces. I thought everybody knew
that they were quite droopy? The radishes were
droopy.
S1: I see, well, she planned to remove them. The
thoughtful gardener thought the flowers charmed
the butterflies.
Fig. 1. Garden Story: Original Version and Monologue/Dialog Generation.
Highlighted areas indicate examples of the same information.
We build on the publicly available PersonaBank corpus 1, which provides us
with the deep story representation and a lexico-syntactic representation of its
monologic retelling [14]. PersonaBank consists of a corpus of monologic personal
narratives from the ICWSM Spinn3r Corpus [6] that are annotated with a deep
story representation called a story intention graph [13]. After annotation,
the stories are run through the est system to generate corresponding deep lin-
guistic structure representations. m2d then takes these representations as input
and creates dialog with different character voices. We identify several sto-
ries by hand as good candidates for dialogic tellings because they describe events
or experiences that two people could have experienced together.
Our primary hypothesis is H1: Dialogic tellings of stories are more engaging
than monologic tellings. We also hypothesize that good dialog requires the use
1Available from nlds.soe.ucsc.edu/personabank

M2D: Monolog to Dialog Generation for Conversational Story Telling 3
of narratological variations such as direct speech, first person, and focalization
[14]. Moreover, once utterances are rendered as first-person with direct speech,
then character voice becomes relevant, because it does not make sense for all
the characters and the narrator to talk in the same voice. Thus our primary
hypothesis H1 entails two additional hypotheses H2 and H3:
H2: Narratological variations such as direct speech, first person, and focal-
ization will affect a readers engagement with a story.
H3: Personality-based variation is a key aspect of expressive variation in
storytelling, both for narrators and story characters. Changes in narrator
or character voice may affect empathy for particular characters, as well as
engagement and memory for a story.
Our approach to creating different character voices is based on the Big Five
theory of personality [1, 9]. It provides a useful level of abstraction (e.g., ex-
traverted vs. introverted characters) that helps to generate language and to
guide the integration of verbal and nonverbal behaviors [16, 21, 11].
To the best of our knowledge, our work is the first to develop and evaluate
algorithms for automatically generating different dialogic tellings of a story from
a deep story representation, and the first to evaluate the utility and effect of
parameterizing the style of speaker voices (personality) while telling the story.
2 Background and Motivation
Stories can be told in either dialog or as a monolog, and in many natural settings
storytelling is conversational [4]. Hypothesis H1 posits that dialogic tellings of
stories will be more engaging than monologic tellings. In storytelling and at least
some educational settings, dialogs have cognitive advantages over monologs for
learning and memory. Students learn better from a verbally interactive agent
than from reading text, and they also learned better when they interacted with
the agent with a personalized dialog (whether spoken or written) than a non-
personalized monolog [20]. Our experiments compare different instances of the di-
alog, e.g. to test whether more realistic conversational exchanges affects whether
people become immersed in the story and affected by it.
Previous work supports H2, claiming that direct, first-person speech in-
creases stories’ drama and memorability [34, 37]. Even when a story is told as a
monolog or with third person narration, dialog is an essential part of stortelling:
in one study of 7 books, between 40% and 60% of the sentences were dialog [7].
In general narratives are mentally simulated by readers [35], but readers also
enact a protagonist’s speech according to her speech style, reading more slowly
for a slow-speaking protagonist and more quickly for a fast-speaking protagonist,
both out-loud and silently [43]. However, the speech simulation only occurred
for direct quotation (e.g. She said “Yeah, it rained”), not indirect quotation (e.g.
She said that it had rained). Only direct quotations activate voice-related parts
of the brain [43], as they create a more vivid experience, because they express
enactments of previous events, whereas indirect quotations describe events [42].

4 Authors Suppressed Due to Excessive Length
Several previous studies also suggest H3, that personality-based variation is
a key aspect of storytelling, both for narrators and story characters. Personality
traits have been shown to affect how people tell stories as well as their choices
of stories to tell [17]. And people also spontaneously encode trait inferences
from everyday life when experiencing narratives, and they derive trait-based
explanations of character’s behavior [30, 32]. Readers use these trait inferences to
make predictions about story outcomes and prefer outcomes that are congruent
with trait-based models [30]. The finding that the behavior of the story-teller is
affected by the personality of both the teller and the listener also motivates our
algorithms for monolog to dialog generation [38, 39]. Content allocation should be
controlled by the personality of the storyteller (e.g. enabling extraverted agents
to be more verbose than introverted agents).
Previous work on generation for fictional domains has typically combined
story and discourse, focusing on the generation of story events and then
using a direct text realization strategy to report those events [18]. This approach
cannot support generation of different tellings of a story [23]. Previous work
on generating textual dialog from monolog suggests the utility of adding extra
interactive elements (dialog interaction) to storytelling and some strategies for
doing so [2, 36, 25]. In addition, expository or persuasive content rendered as
dialog is more persuasive and memorable [41, 26, 27]. None of this previous work
attempts to generate dialogic storytelling from original monologic content.
3 M2D: Monolog-to-Dialog Generation
Scheherazade
ES-Translator
M2D: Monolog
to Dialog RealPro
Corpus of Personal Narratives/Stories
Monolog
Input
Dialog
Output
DsyntS Altered DsyntS
Fig. 2. m2d Pipeline Architecture.
Figure 2 illustrates the architecture
of m2d. The est framework produces
a story annotated by scheherazade
as a list of Deep Syntactic Structures
(DsyntS). DsyntS, the input format
for the surface realizer RealPro [12,
19], is a dependency-tree structure
where each node contains the lexical
information for the important words
in a sentence. Each sentence in the
story is represented as a DsyntS.
m2d converts a story (as a list of DsyntS) into different versions of a two-
speaker dialog using a parameterizable framework. The input parameters control,
for each speaker, the allocation of content, the usage of questions of different
forms, and the usage of various pragmatic markers (Table 1). We describe the
m2d parameters in more details below.
Content allocation: We allocate the content of the original story between the
two speakers using a content-allocation parameter that ranges from 0 to 1. A
value of .5 means that the content is equally split between 2 speakers. This is
motivated by the fact that, for example, extraverted speakers typically provide
more content than intraverted speakers [16,38].

M2D: Monolog to Dialog Generation for Conversational Story Telling 5
Table 1. Dialog Conversion Parameters
Parameter Example
Aggregation
Merge short sents The garden was swampy, and not productive.
Split long sents The garden was very swampy because it rained. The garden is very large,
and has lots of plants.
Coreference
Pronominalize The gardener likes to eat apples from his orchard. They are red.
Pragmatic Markers
Emphasizer great Great, the garden was swampy.
Downer kind of The garden was kind of swampy.
Acknowledgment yeah Yeah, the garden was swampy.
Repetition S1: The garden was swampy.
S2: Yeah, the garden was swampy.
Paraphrase S0
2: Right, the garden was boggy.
Interactions
Affirm Adjective S1: The red apples were tasty and −−−
S2: Just delicious, really.
S1: Yeah, and the gardener ate them.
Correct Inaccuracies S2: The garden was not productive and −−−.
S1: I don’t think that’s quite right, actually. I think the garden was pro-
ductive.
Questions
Provoking I don’t really remember this part, can you tell it?
With Answer S1: How was the garden?
S2: The garden was swampy.
Tag The garden was swampy, wasn’t it?
Character and Property Database: We use the original source material for
the story to infer information about actors, items, groups, and other properties of
the story, using the information specified in the DsyntS. We create actor objects
for each character and track changes in the actor states as the story proceeds, as
well as changes in basic properties such as their their body parts and possessions.
Aggregation and Deaggregation: We break apart long DsyntS into smaller
DsyntS, and then check where we can merge small and/or repetitious DsyntS.
We believe that deaggregation will improve our dialogs overall clarity while ag-
gregation will make our content feel more connected [15].
Fig. 3. The DsyntS tree for The man
ran to the big store.
Content Elaboration: In natural di-
alog, speakers often repeat or partially
paraphrase each other, repeating the same
content in multiple ways. This can be a
key part of entrainment. Speakers may
also ask each other questions thereby set-
ting up frames for interaction [38]. In our
framework, this involves duplicating con-
tent in a single DsyntS by either 1) gener-
ating a question/answer pair from it and
allocating the content across speakers, or
2) duplicating it and then generating paraphrases or repetitions across speakers.
Questions are generated by performing a series of pruning operations based on
the class of the selected node and the relationship with its parent and siblings.
For example, if store in Figure 3 is selected, we identify this node as our ques-
tion. The class of a node indicates the rules our system must follow when making
deletions. Since store is a noun we prune away all of the attr siblings that

6 Authors Suppressed Due to Excessive Length
modify it. By noticing that it is part of a prepositional phrase, we are able to
delete store and use to as our question, generating The man ran where?.
Content Extrapolation: We make use of the deep underlying story represen-
tation and the actor database to make inferences explicit that are not actually
part of the original discourse. For example, the actor database tracks aspects
of a character’s state. By using known antonyms of the adjective defining the
current state, we can insert content for state changes, i.e. the alteration from
the fox is happy to now, the fox is sad, where the fox is the actor and happiness
is one of his states. This also allows us to introduce new dialogic interactions
by having one speaker ask the other about the state of an actor, or make one
speaker say something incorrect which allows the second speaker to contradict
them: The fox was hungry followed by No, he wasn’t hungry, he was just greedy.
Pragmatic Markers: We can also insert pragmatic markers and tag questions
as described in Table 1. Particular syntactic constraints are specified for each
pragmatic marker that controls whether the marker can be inserted at all [16].
The frequency and type of insertions are controlled by values in the input param-
eter file. Some parameters are grouped by default into sets that allow them to
be used interchangeably, such as downtoners or emphasizers. To provide us more
control over the variability of generated variants, specific markers which are by
default unrelated can be packaged together and share a distributed frequency
limit. Due to their simplistic nature and low number of constraints, pragmatic
markers prove to be a reliable source of variation in the systems output.
Lexical Choice: We can also replace a word with one of its synonyms. This
can be driven simply by a desire for variability, or by lexical choice parameters
such as word frequency or word length.
Morphosyntactic Postprocessing: The final postprocessing phase forms con-
tractions and possessives and corrects known grammatical errors.
The results of the m2d processor are then given as input to RealPro [12],
an off-the-shelf surface text realizer. RealPro is responsible for enforcing English
grammar rules, morphology, correct punctuation, and inserting functional words
in order to produce natural and grammatical utterances.
4 Evaluation Experiments
We assume H1 on the basis of previous experimental work, and test H2 and H3.
Our experiments aim to: (1) establish whether and to what degree the m2d en-
gine produces natural dialogs; (2) determine how the use of different parameters
affect the user’s engagement with the story and the user’s perceptions of the
naturalness of the dialog; and (3) test whether users perceive personality dif-
ferences that are generated using personality models inspired by previous work.
All experimental participants are pre-qualified Amazon Mechanical Turkers to
guarantee that they provide detailed and thoughtful comments.
We test users’ perceptions of naturalness and engagement using two stories:
the Garden story (Figure 1) and the Squirrel story (Figure 4). For each story,
we generate three different dialogic versions with varying features:

M2D: Monolog to Dialog Generation for Conversational Story Telling 7
m2d-est renders the output from est as a dialog by allocating the content
equally to the two speakers. No variations of sentences are introduced.
m2d-basic consists of transformations required to produce a minimally natu-
ral dialog. First we apply pronominalization to replace nouns with their
pronominal forms when telling the story in the third person. We then manip-
ulate sentence length by breaking very long sentences into shorter ones, or
by combining repetitious short sentences into one sentence. This is motivated
by the fact that utterances in dialog tend to be less formal and use less com-
plex syntactic structures [5]. The last transformation is morphosyntactic
postprocessing as described in Section 3.
m2d-chatty adds interactive features to m2d-basic such as the insertion of
pragmatic markers (acknowledgements, disfluencies, hedges) and question-
answer generation (Table 1).
Squirrel Story: Monolog generated by est
I placed the steely bowl on the deck in order for Benjamin to drink the bowl’s water. The steely
bowl was popular. The birds drank the bowl’s water. The birds bathed themselves in the steely
bowl. The birds organized themselves on the deck’s railing in order for the birds to wait. The
squirrels drank the bowl’s water. The squirrel approached the steely bowl. The crazy squirrel was
startled because the squirrel saw the squirrel’s reflection. The crazy squirrel leaped because the
squirrel was startled. The crazy squirrel fell over the deck’s railing because the squirrel leaped
because the squirrel was startled. The crazy squirrel held the deck’s railing with the squirrel’s
paw. The squirrel’s paw slipped off the deck’s railing. The crazy squirrel fell.
Squirrel Story: Dialog generated by m2d
S1: I placed the steely bowl on the deck. Benjamin drinks its water.
S2: The steely bowl was very popular, technically. Typical. The birds drank its water. They bathed
themselves in the steely bowl! The birds organized themselves on the deck’s railing in order for
them to wait.
S1: Yeah, the birds organized themselves on the deck’s railing in order for them to wait. The
squirrels drank it’s water. The squirrel approached the steely bowl.
S2: Technically, it was startled. I am delighted to say that the squirrel saw its reflection. It
jumped, actually and was startled. Basically, the squirrel was literally startled and fell over the
deck’s railing. It leaped because it was startled.
S1: The crazy squirrel held the deck’s railing with its paw.
S2: Its paw slipped off its railing pal. The particularly crazy squirrel fell mate.
Fig. 4. Squirrel Story: Monolog/Dialog Generation.
Each pairwise comparison is a Human Intelligence Task (HIT; a question that
needs an answer), yielding 6 different HITs. We used 5 annotators (Turkers) per
HIT to rate the levels of engagement/naturalness on a scale of 1-5, followed by
detailed comments justifying their ratings.
We create several subsets of features that work well together and recursively
apply random feature insertion to create many different output generations.
These subsets include the types of questions that can be asked, different speaker
interactions, content polarity with repetition options, pragmatic markers, and
lexical choice options. A restriction is imposed on each of the subgroups, indicat-
ing the maximum number of parameters that can be enabled from the associated
subgroup. This results in different styles of speaker depending on which subset
of features is chosen. A speaker who has a high number of questions along with
hedge pragmatic markers will seem more inquisitive, while a speaker who just
repeats what the other speaker says may appear to have less credibility than the
other speaker. We plan to explore particular feature groupings in future work to
identify specific dialogic features that create a strong perception of personality.

8 Authors Suppressed Due to Excessive Length
4.1 M2D-EST vs. -Basic vs. -Chatty
The perceptions of engagement given different versions of the dialogic story is
shown in Figure 5. A paired t-test comparing m2d-chatty to m2d-est shows
that increasing the number of appropriate features makes the dialog more en-
gaging (p = .04, df= 9). However there are no statistically significant differences
between m2d-basic and m2d-est, or between m2d-basic and m2d-chatty.
Comments by Turkers suggest that the m2d-chatty speakers have more per-
sonality because they use many different pragmatic markers, such as questions
and other dialogically oriented features.
Fig. 5. Mean scores for
engagement.
Fig. 6. Mean scores for
naturalness.
The perception of nat-
uralness across the same
set of dialogic stories is
shown in Figure 6. It shows
that m2d-basic was rated
higher than m2d-est, and
their paired t-test shows
that m2d-basic (inclusion
of pronouns and agg- and deaggregation) has a positive impact on the natural-
ness of a dialog (p = .0016, df = 8). On the other hand, m2d-basic is preferred
over m2d-chatty, where the use of pragmatic markers in m2d-chatty was
often noted as unnatural.
4.2 Personality Models
A second experiment creates a version of m2d called m2d-personality which
tests whether users perceive the personality that m2d-personality intends to
manifest. We use 4 different stories from the PersonaBank corpus [13] and create
introverted and extroverted personality models, partly by drawing on features
from previous work on generating personality [16].
We use a number of new dialogic features in our personality models that
increase the level of interactivity and entrainment, such as asking the other
speaker questions or entraining on their vocabulary by repeating things that they
have said. Content allocation is also controlled by the personality of the speaker,
so that extraverted agents get to tell more of the content than introverted agents.
We generate two different versions of each dialog, an extroverted and an
introverted speaker (Table 2). Each dialog also has one speaker who uses a default
personality model, neither strongly introverted or extraverted. This allows us to
test whether the perception of the default personality model changes depending
on the personality of the other speaker. We again created HITs for Mechanical
Turk for each variation. The Turkers are asked to indicate which personality best
describes the speaker from among extroverted, introverted, or none, and then
explain their choices with detailed comments. The results are shown in Table 3,
where Turkers correctly identified the personality that m2d-personality aimed
to manifest 88% of the time.

M2D: Monolog to Dialog Generation for Conversational Story Telling 9
Table 3. Personality Judgments
Extro Intro None
Extro 16 0 0
Intro 3 12 1
Table 4. Default Personality Judgments
Extro Intro None
Other is Extro 1 7 8
Other is Intro 8 1 7
Table 2. Feature Frequency for Extra. vs. Intro.
Not all lexical instantiations of a feature are listed.
Parameter Extra. Intro.
Content allocation
Content density high low
Pragmatic markers
Adjective softeners low high
Exclamation high low
Tag Questions high low
Acknowledgments: Yeah, oh God high low
Acknowledgments: I see, well, right low high
Downtoners: Sort of, rather, quite, pretty low high
Uncertainty: I guess, I think, I suppose low high
Filled pauses: Err..., Mmhm... low high
Emphasizers: Really, basically, technically high low
In-group Markers: Buddy, pal high low
Content elaboration
Questions: Ask & let other spkr answer high low
Questions: Rhetorical, request confirmation low high
Paraphrase high low
Repetition low high
Interactions: Affirm adjective high low
Interactions: Corrections high low
Lexical choice
Vocabulary size high low
Word length high low
Turkers’ comments noted
the differential use of prag-
matic markers, content al-
location, asking questions,
and vocabulary and punc-
tuation. The extroverted
character was viewed as
more dominant, engaging,
excited, and confident. These
traits were tied to the
features used: exclamation
marks, questions asked, ex-
changes between speakers,
and pragmatic markers (e.g.,
basically,actually).
The introverted char-
acter was generally timid,
hesitant, and keeps their
thoughts to themselves. Turkers noticed that the introverted speaker was al-
located less content, the tendency to repeat what has already been said, and the
use of different pragmatic markers (e.g. kind of,I guess,Mhmm,Err...).
Table 4 shows Turker judgements for the speaker in each dialog who had a
default personality model. In 53% of the trials, our participants picked a per-
sonality other than “none” for the speaker that had the default personality.
Moreover, in 88% of these incorrect assignments, the personality assigned to the
speaker was the opposite of the personality model assigned to the other speaker.
These results imply that when multiple speakers are in a conversation, judge-
ments of personality are relative to the other speaker. For example, an introvert
seems more introverted in the presence of an extravert, or a default personality
may seem introverted in the presence of an extravert.
5 Discussion and Future Work
We hypothesize that dialogic storytelling may produce more engagement in the
listener, and that the capability to render a story as dialog will have many prac-
tical applications (e.g. with gestures [10]. We also hypothesize that expressing
personality in storytelling will be useful and show how it is possible to do this in
the experiments presented here. We described an initial system that can translate
a monologic deep syntactic structure into many different dialogic renderings.
We evaluated different versions of our m2d system. The results indicate that
the perceived levels of engagement for a dialogic storytelling increase propor-

10 Authors Suppressed Due to Excessive Length
tionally with the density of interactive features. Turkers commented that the
use of pragmatic markers, proper pronominalization, questions, and other in-
teractions between speakers added personality to the dialog, making it more
engaging. In a second experiment, we directly test whether Turkers perceive
that different speaker’s personalities in dialog. We compared introvert, extro-
vert, and a speaker with a default personality model. The results show that in
88% of cases the reader correctly identified the personality model assigned to the
speaker. The results show that the content density assigned to each speaker as
well as the choice of pragmatic markers are strong indicators of the personality.
Pragmatic markers that most emphasize speech, or attempt to engage the other
speaker are associated with extroverts, while softeners and disfluencies are asso-
ciated with introverts. Other interactions such as correcting false statements and
asking questions also contribute to the perception of the extroverted personality.
In addition, the perceived personality of the default personality speaker was
affected by the personality of the other speaker. The default personality speaker
was classified as having a personality 53% of the time. In 88% of these misclas-
sifications, the personality assigned to the speaker was the opposite of the other
speaker, suggesting that personality perception is relative in context.
While this experiment focused only on extrovert and introvert, our frame-
work contains other Big-Five personality models that can be explored in the
future. We plan to investigate: 1) the effect of varying feature density on the
perception of a personality model, 2) how personality perception is relative in
context, and 3) the interaction of particular types of content or dialog acts with
perceptions of a storyteller’s character or personality. The pragmatic markers
are seen as unnatural in some cases. We note that our system currently inserts
them probabilistically but do not make intelligent decisions about using them
in pragmatically appropriate situations. We plan to add this capability in the
future. In addition we will explore new parameters that improves the naturalness
and flow of the story.
Acknowledgments
We would like to thank Chung-Ning Chang and Diego Pedro for their roles as
collaborators in the early inception of our system. This research was supported
by NSF IIS CHS #1115742 and award #SC-14-74 from the Nuance Foundation.
References
1. Gordon W. Allport and Henry S. Odbert. Trait names: a psycho-lexical study.
Psychological Monographs, 47(1, Whole No. 211):171–220, 1936.
2. Elisabeth Andr´e, Thomas Rist, Susanne van Mulken, Martin Klesen, and Stephan
Baldes. The automated design of believable dialogues for animated presentation
teams. Embodied conversational agents, pages 220–255, 2000.
3. Mieke Bal and Eve Tavor. Notes on narrative embedding. Poetics Today, pages
41–59, 1981.

M2D: Monolog to Dialog Generation for Conversational Story Telling 11
4. Janet B Bavelas, Linda Coates, and Trudy Johnson. Listeners as co-narrators.
Journal of personality and social psychology, 79(6):941, 2000.
5. D. Biber. Variation across speech and writing. Cambridge Univ Pr, 1991.
6. Kevin Burton, Akshay Java, and Ian Soboroff. The icwsm 2009 spinn3r dataset.
In Proc. of the Annual Conf. on Weblogs and Social Media (ICWSM), 2009.
7. Pieter De Haan. More on the language of dialogue in fiction. ICAME JOURNAL,
20:23–40, 1996.
8. G´erard Genette. Nouveau discours du r´ecit.´
Ed. du Seuil, 1983.
9. Lewis R. Goldberg. An alternative “description of personality”: The Big-Five factor
structure. Journal of Personality and Social Psychology, 59:1216–1229, 1990.
10. Zhichao Hu, Michelle Dick, Chung-Ning Chang, Kevin Bowden, Michael Neff, Jean
Fox Tree, and Marilyn Walker. A Corpus of Gesture-Annotated Dialogues for
Monologue-to-Dialogue Generation from Personal Narratives. Proc. of the Tenth
International Conf. on Language Resources and Evaluation (LREC 2016), 2016.
11. Zhichao Hu, Marilyn Walker, Michael Neff, and Jean E. Fox Tree. Storytelling
agents with personality and adaptivity. Intelligent Virtual Agents, page 30, 2015.
12. Benoit Lavoie and Owen Rambow. A fast and portable realizer for text generation
systems. In Proc. of the Third Conf. on Applied Natural Language Processing,
ANLP97, pages 265–268, 1997.
13. Stephanie Lukin, Kevin Bowden, Casey Barackman, and Marilyn Walker. Person-
abank: A corpus of personal narratives and their story intention graphs.
14. Stephanie Lukin and Marilyn Walker. Narrative variations in a virtual storyteller.
Intelligent Virtual Agents, page 30, 2015.
15. Stephanie M Lukin, Lena I Reed, and Marilyn Walker. Generating sentence plan-
ning variations for story telling. In 16th Annual Meeting of the Special Interest
Group on Discourse and Dialogue, page 188, 2015.
16. F. Mairesse and M.A. Walker. Towards personality-based user adaptation: psycho-
logically informed stylistic language generation. User Modeling and User-Adapted
Interaction, pages 1–52, 2010.
17. D.P. McAdams, N.A. Anyidoho, C. Brown, Y.T. Huang, B. Kaplan, and M.A.
Machado. Traits and stories: Links between dispositional and narrative features of
personality. Journal of Personality, 72(4):761–784, 2004.
18. J Meehan. The metanovel: writing stories by computer, pages 197–258. 1976.
19. Igor A. Melˇcuk. Dependency Syntax: Theory and Practice. SUNY, Albany, New
York, 1988.
20. Roxana Moreno, Richard Mayer, and James Lester. Life-like pedagogical agents
in constructivist multimedia environments: Cognitive consequences of their inter-
action. In World Conf. on Educational Media and Technology, number 1, pages
776–781, 2000.
21. M. Neff, Y. Wang, R. Abbott, and M. Walker. Evaluating the effect of gesture and
language on personality perception in conversational agents. In Intelligent Virtual
Agents, pages 222–235. Springer, 2010.
22. Monisha Pasupathi and Timothy Hoyt. The development of narrative identity in
late adolescence and emergent adulthood: The continued importance of listeners.
Developmental Psychology, 45(2):558–574, 2009.
23. Federico Peinado and Pablo Gerv´as. Evaluation of automatic generation of ba-
sic stories. New Generation Computing, Special issue: Computational Creativity,
24(3):289–302, 2006.
24. James W Pennebaker and Janel D Seagal. Forming a story: The health benefits of
narrative. Journal of clinical psychology, 55(10):1243–1254, 1999.

12 Authors Suppressed Due to Excessive Length
25. P. Piwek and S. Stoyanchev. Generating expository dialogue from monologue:
motivation, corpus and preliminary rules. In Human Language Technologies: The
2010 Annual Conf. of the North American Chapter of the Association for Com-
putational Linguistics, pages 333–336. Association for Computational Linguistics,
2010.
26. Paul Piwek, Hugo Hernault, Helmut Prendinger, and Mitsuru Ishizuka. T2d: Gen-
erating dialogues between virtual agents automatically from text. In Intelligent
Virtual Agents, pages 161–174. Springer, 2007.
27. Paul Piwek and Svetlana Stoyanchev. Data-oriented monologue-to-dialogue gen-
eration. In Proc. of the 49th Annual Meeting of the Association for Computational
Linguistics: Human Language Technologies: short papers-Volume 2, pages 242–247.
Association for Computational Linguistics, 2011.
28. Gerald Prince. A Grammar of Stories: An Introduction. Number 13. Walter de
Gruyter, 1973.
29. Vladimir Propp. Morphology of the Folk Tale. University of Texas Press, second
edition, 1969. Trans. Laurence Scott. Originally published 1928.
30. D.N. Rapp, R.J. Gerrig, and D.A. Prentice. Readers’ Trait-Based Models of Char-
acters in Narrative Comprehension. Journal of Memory and Language, 45(4):737–
750, 2001.
31. Elena Rishes, Stephanie M Lukin, David K Elson, and Marilyn A Walker. Generat-
ing different story tellings from semantic representations of narrative. In Interactive
Storytelling, pages 192–204. Springer International Publishing, 2013.
32. L. Ross. The intuitive psychologist and his shortcomings: Distortions in the attri-
bution process. Advances in experimental social psychology, 10:173–220, 1977.
33. Kimiko Ryokai, Catherine Vaucelle, and Justine Cassell. Virtual peers as part-
ners in storytelling and literacy learning. Journal of Computer Assisted Learning,
19(2):195–208, 2003.
34. Deborah Schiffrin. Tense variation in narrative. Language, pages 45–62, 1981.
35. Nicole K Speer, Jeremy R Reynolds, Khena M Swallow, and Jeffrey M Zacks.
Reading stories activates neural representations of visual and motor experiences.
Psychological Science, 20(8):989–999, 2009.
36. S. Stoyanchev and P. Piwek. Harvesting re-usable high-level rules for expository
dialogue generation. In Proc. of the 6th International Natural Language Generation
Conf., pages 145–154. Association for Computational Linguistics, 2010.
37. Deborah Tannen. Talking voices : repetition, dialogue, and imagery in conversa-
tional discourse. CUP, 1989.
38. A. Thorne. The press of personality: A study of conversations between introverts
and extraverts. Journal of Personality and Social Psychology, 53(4):718, 1987.
39. Avril Thorne, Neill Korobov, and Elizabeth M Morgan. Channeling identity: A
study of storytelling in conversations between introverted and extraverted friends.
Journal of research in personality, 41(5):1008–1031, 2007.
40. Avril Thorne, Kate C McLean, and Amy M Lawrence. When remembering is
not enough: Reflecting on self-defining memories in late adolescence. Journal of
personality, 72(3):513–542, 2004.
41. Kees Van Deemter, Brigitte Krenn, Paul Piwek, Martin Klesen, Marc Schr¨oder, and
Stefan Baumann. Fully generated scripted dialogue for embodied agents. Artificial
Intelligence, 172(10):1219–1244, 2008.
42. E. Wade and H.H. Clark. Reproduction and demonstration in quotations. Journal
of Memory and Language, 32(6):805 – 819, 1993.
43. Bo Yao and Christoph Scheepers. Contextual modulation of reading rate for direct
versus indirect speech quotations. Cognition, 121(3):447–453, 2011.