=Paper=
{{Paper
|id=Vol-2342/paper10
|storemode=property
|title=Automatic Recognition of Narrative Drama Units: A Structured Learning Approach
|pdfUrl=https://ceur-ws.org/Vol-2342/paper10.pdf
|volume=Vol-2342
|authors=Danilo Croce,Roberto Basili,Vincenzo Lombardo,Eleonora Ceccaldi
|dblpUrl=https://dblp.org/rec/conf/ecir/CroceBLC19
}}
==Automatic Recognition of Narrative Drama Units: A Structured Learning Approach==
https://ceur-ws.org/Vol-2342/paper10.pdf
Automatic Recognition of Narrative Drama units: a
structured learning approach
Danilo Croce Roberto Basili
Dept. of Enterprise Engineering Dept. of Enterprise Engineering
University of Roma, Tor Vergata (Italy) University of Roma, Tor Vergata (Italy)
croce@info.uniroma2.it basili@info.uniroma2.it
Vincenzo Lombardo Eleonora Ceccaldi
Dept. of Informatics, DIBRIS
University of Torino, (Italy) University of Genova, (Italy)
vincenzo.lombardo@unito.it eleonoraceccaldi@gmail.com
Abstract
Drama is a story told through the live actions of characters; dramatic
writing is characterized by aspects that are central to identify, interpret,
and relate the different elements of a story. The Drammar ontology has
been proposed to represent core dramatic qualities of a dramatic text,
namely Actions, Agents, Scenes and Conflicts, evoked by individual
text units. The automatic identification of such elements in a drama is
the first step in the recognition of their evolution, both at coarse and
fine grain text level. In this paper, we address the issue of segmenta-
tion, that is, the partition of the drama into meaningful unit sequences
We study the role of editorial as well as content–based text properties,
without relying on deep ontological relations. We propose a genera-
tive inductive machine learning framework, combining Hidden Markov
models and SVM and discuss the role of event information (thus in-
volving agents and actions) at the lexical and grammatical level.
1 Introduction
Drama is a story told through the live actions of characters. The Drammar ontology [LDP18, DLPar] identifies
the core dramatic qualities of a dramatic text, namely Actions, Agents, Units/Scenes, and Conflicts, implicitly
evoked by the dramatic text, as claimed by the scientific literature on drama analysis.
Drama relies on an internal coherence and a rich set of eventualities, related to the interactions among
characters and the insurgence and resolution of conflicts. Dramas are very well structured. As a running
example, we address the incipit of Anton Chekhov’s “The Cherry Orchard ”, in its English translation [Che17]:
A room which is still called the nursery. One of the doors leads into ANYA’S room. It is close on sunrise. ...
DUNYASHA comes in with a candle, and LOPAKHIN with a book in his hand.
LOPAKHIN. The train’s arrived, thank God. What’s the time?
Copyright c 2019 for the individual papers by the paper’s authors. Copying permitted for private and academic purposes. This
volume is published and copyrighted by its editors.
In: A. Jorge, R. Campos, A. Jatowt, S. Bhatia (eds.): Proceedings of the Text2StoryIR’19 Workshop, Cologne, Germany, 14-April-
2019, published at http://ceur-ws.org
DUNYASHA. It will soon be two. [Blows out candle] It is light already.
Schematically:
• individual utterances are denoted by the correspondent acting characters;
• some editorial notes (in italics) interleave with the spoken parts where the authors suggest the environment
changes or specific happenings;
• a strict separation between spoken and editorial fragments is imposed.
Our current research objective aims at supporting the automatic annotation of a drama, able to outline the
evolution of the dramatic elements above through discrete events.
1.1 Events in narrative texts
Following the observer point of view, [SZR07] propose the following definitions for the events:
• An Event is ”a segment of time at a given location, that is conceived by an observer to have a beginning
and an end; granularity of events can go from a second or less to tens of minutes”.
• An Event model is ”an actively maintained representation of the current event, which is updated at
perceptual event boundaries”.
• The Event segmentation is ”the perceptual and cognitive process by which a continuous activity is
segmented into meaningful events”.
From the psychological literature, we have that readers structure a narrative text into a series of events in order
to understand and remember the text (cf. the experiments of [ZT01, ZS07, ZSSM10]). Events are coded at
clause level. Relevant information for the narrative coding includes, e.g. [SZR07]:
• Time and Space information (as the presence of spatial changes, e.g., moving from one room to another
inside a house can be meaningful);
• Objects, given the interaction of characters with elements of a scene;
• change of Character, revealed by the changes of the subject of a clause;
• Causes (causal relationship over activities) and Goals (new goal-directed activities), to be coded as core
dimensions of Events.
Usually, clauses are also coded for terminal punctuation (e.g., periods and question marks) and non terminal
punctuation (e.g., commas and semicolons). As the annotation of such dramatic aspects is time consuming, we
aim at automatizing it, relying upon the lexical, grammatical, and editorial information, expressed by individual
clauses. In this way, events can be recognized and properly segmented along the dramatic text. We will refer
hereafter this process as the event segmentation.
1.2 Event Segmentation: Related Work
Event segmentation is a task traditionally tackled in NLP according to sentence boundary detection methods
(e.g. [Hea94], [SDDK11]) or cohesion-based clustering models (e.g. [Cho00]). Text segmentation methods usually
search for the set of segments in a text that optimize some form of coherence of the content. Word usage is modeled
in TextTiling [Hea94] for each sentence in a sequence and the two sides of a potential boundary are selected
when large lexical difference is found. Prosodic features and lexical features are taken into account to model
discourse as in the Hidden Markov Model segmentation proposed in [YXX+ 16]. The lexical connectivity strength
between two adjacent fragments of a text is used as hint in DivSeg ([SDDK11]). Unsupervised approaches are
based on probabilistic models (e.g. [Hea94], C99 [Cho00] or the DotPlotting algorithm [Rey94]) or agglomerative
clustering [Yaa99]). In the former group, terms frequencies are used to identify topical segments (dense dot
clouds on the graphic). e.g. DotPlotting [Rey94]. In the latter group, dendrograms are induced over paragraphs
and transformed into a hierarchical segmentation [Yaa99]. Lexical chains methods are applied in an unsupervised
manner as they exploit semantic lexicons to model word associations and semantic relations. In these methods,
a chain links multiple occurrences of a term in the document: it is considered broken when there are too many
sentences between two occurrences of a term. The Segmenter system ([KKM98]) detects such broken points
across a document according to possibly multiple chains. Some of the methods use lexical resources or forms
of ontological similarity to model similarity metrics between text blocks (sentences or paragraphs), based on
semantic information (e.g. recognized named entity in the text). Wordnet or Wikipedia-based methods have
been proposed to define semantic similarity metrics between text units. Recently, deep learning methods have
been applied to Text Segmentation, specifically to the Topic-based segmentation task. In particular [LSJ18]
presents an end-to-end segmentation model: first, a bidirectional recurrent neural network is used to encode
input text sequences, and then, another recurrent neural network is used together with a pointer network to
select text boundaries in the input sequence. Although very appealing, since it does not require hand-crafted
features definition, this method requires a significant amount of training material, made of several hundreds of
annotated documents.
2 A structured learning approach to drama segmentation
In line with most of the above event segmentation approaches, we will rely on a machine learning perspective by
assuming a set of textual resources as the triggering observations:
• L is a set of fully annotated drama fragments, whose segments are completely known (e.g., the example
units 4 to 6 in Checkov’s The Cherry Orchard reported in the paper Appendix or the nunnery scene from
Shakespeare’s Hamlet[LPD16], respectively);
• OL is a very small scale corpus, made of fragments from the possibly partially annotated opera (e.g., the
complete annotated dramas “The Cherry Orchard” or “Hamlet”, respectively, though they are usually
neither segmented, nor annotated);
• OA (L) is a large scale corpus of unannotated texts of the same author (e.g., all of Checkov’s plays or
Shakespeare’s plays, respectively);
• OE (A(L)) is a comprehensive corpus of the drama works of the same epoch (e.g., Contemporary play or
Elizabethan theatre plays, respectively).
So, we rely on the chain L ⊂ OL ⊂ OA (L) ⊂ OE (A(L)).
We propose an integration of unsupervised and supervised learning processes acting: our first attempt is to
use the comprehensive OE (A(L)) to generate a lexical resource focused on the work and author style: according
to unsupervised methods, such as [MCCD13], we can rely on word embedding for a large scale dictionary of
lexical items: these generalize lexical semantics within the underlying targeted text genre. The proposal is to
inject this information into the supervised steps that address the labeled material L, in order to fully label the
entire work OL in an accurate manner. Annotated examples in L are the basic source of information for the
segmentation stage.
Hereafter, we concentrate on the variety of lexical, grammatical and aspectual features (e.g. the mode and
transitivity of a number of verbs involved in the dramatic action), suitably exploited for training a sequence
labeling component over OL . We propose a structured learning paradigm based on independent kernels for
training SVMs over L ([STC04]) and apply them within a Markovian modeling, isomorphic to HMM. The major
steps are thus:
• (PreTraining) Use OE (A(L)) to acquire lexical information in the form of a neural language model (in line
with [MCCD13]), expressing general semantic properties of individual words. A specific treatment of some
classes of words is here applied. For example, character names (e.g. Dunyasha and Lopakhin in The Cherry
Orchard or Hamlet and Ophelia in Hamlet). This is a standard a-priori information for a drama that is
mapped into the category label Character, in order to minimize sparsity.
• (Feature Modeling and Extraction) Feature extraction is applied to derive textual, editorial and nar-
rative features, as discussed in Section 2.2
• (Model Optimization) Then, a structured Machine Learning model is applied to achieve segmentation as
an IOB-like sentence labeling process, in order to organize sentences in units and hierarchies of scenes. The
adopted algorithm is known as SVM-HMM ([TJHA05], adopted in [CB11, BCV+ 16]).
2.1 A Markovian Support Vector Machine
The aim of a Markovian formulation of SVM is to make the classification of a input example xi ∈ Rn (belonging
to a sequence of examples) dependent on the labels assigned to the previous elements in a history of length m,
i.e., xi−m , . . . , xi−1 . In our classification task, a drama is a sequence of utterances x = (x1 , . . . , xs ), each of them
representing the example xi , i.e., the specific i-th paragraph. Given the corresponding sequence of expected
labels y = (y1 , . . . , ys ), a sequence of m step-specific labels (from a dictionary of d symbols) can be retrieved, in
the form yi−m , . . . , yi−1 . In our machine learning setting, labels are related to the Segmentation task: we will
thus adopt the IOB notation so that each element in the drama will be associated to the label B if it is at the
Beginning of a Unit, I if it is Inside it, O if it is Out of the Unit itself. In order to make the classification of xi also
dependent on the previous decisions, we augment the feature vector of xi by introducing a projection function
ψm (xi ) ∈ Rmd that associates each example with a md−dimensional feature vector where each dimension set to
1 corresponds to the presence of one of the d possible labels observed in a history of length m, i.e. m steps before
the target element xi .
In order to apply a SVM, a projection function φm (·) can be defined to consider both the observations xi
and the transitions ψm (xi ) by concatenating the two representations as follows: φm (xi ) = xi || ψm (xi ) with
φm (xi ) ∈ Rn+md . Notice that the symbol || here denotes the vector concatenation, so that ψm (xi ) does not
interfere with the original feature space, where xi lies. Kernel-based methods can be applied in order to model
meaningful representation spaces, encoding both the feature representing individual examples together with the
information about the transitions. According to kernel-based learning [STC04], we can define a kernel function
Km (xi , zj ) between a generic item of a sequence xi and another generic item zj from the same or a different
sequence, parametric in the history length m. It surrogates the dot product between φm (·) such that:
Km (xi , zj ) = φm (xi )φm (zj ) = K obs (xi , zj ) + K tr ψm (xi ), ψm (zj )
We define a kernel that is the linear combination of two further kernels: K obs operating over the individual
examples xi and a K tr operating over the feature vectors encoding the involved transitions. It is worth noticing
that K obs neither depends on the position nor on the context of individual examples, in line with Markov
assumption that characterizes a large class of these generative models, e.g. HMM. For simplicity, we define K tr
as a linear kernel between input instances, i.e. a dot-product in the space generated by ψm (·):
Km (xi , zj ) = K obs (xi , xj ) + ψm (xi )ψm (zj )
At training time, we use the kernel-based SVM in a One-Vs-All schema over the feature space derived by
Km (·, ·). The learning process provides a family of classification functions f (xi ; m) ⊂ Rn+md × Rd , which asso-
ciate each xi to a distribution of scores with respect to the different d labels, depending on the context size m.
At classification time, all possible sequences y ∈ Y + should be considered in order toP determine the best labeling
ŷ, where m is the size of the history used to enrich xi , that is: ŷ = arg maxy∈Y + { i=1...m f (xi ; m)}
In order to reduce the computational cost, a Viterbi-like decoding algorithm is adopted1 as described in Fig.
1. The next section defines the kernel function K obs applied to specific turns in the drama.
2.2 Modeling dramatic properties as ML features
Three types of kernels are applied for different types of features.
Lexical features include sentence embeddings as linear combinations of individual word embeddings, grammat-
ical patterns, such as verb-objects or subject-verb pairs, POS n-grams (n=3) and, finally, sentence properties
such as length and complexity (e.g. number of different active mode verbs).
Narrative features are strictly dependent on the narrative structure and express possible Characters and
Actions in a turn. Named Entity Recognition is first run on the individual utterances, to capture character
mentions. A narrative vector including the acting character (e.g. LOPAKHIN in line 0036 or 0038 in the Ap-
pendix) as well as all the other recently mentioned characters (e.g. LOPAKHIN and DUNYASHA in the editorial
note at line 0042). Individual features modeling the number of mentioned or recently mentioned characters for
each turn will be adopted. An aging mechanism defines lower scores for no longer mentioned characters. Fi-
nally, narrative features denoting the Actions mentioned in a turn will be adopted in order to account for the
interaction (and possible conflicts) in an explicit way. Examples are motion verbs such as to come, to go, social
verbs, such as to meet (see LOPAKHIN in unit 0040) or even emotional verbs (e.g. or to faint as in unit 0041.
1 When applying f (x ; m) the classification scores are normalized through a softmax function and probability scores are derived.
i
Figure 1: The overall sequence labeling architecture for event segmentation.
Specific dictionaries of English verbs and their nominalizations will be used here to denote narratively interesting
Actions.
Editorial features will depend on the material that includes the author’s suggestions in the environment (see,
for example, the sentence “A room which is still called the nursery” in the incipit). In this case, a representation
that is similar to the one for the lexical features for individual acting turns is adopted, but the editorial material
will be expressed through a separated vector, in order to play an independent role.
Table 1: Performance scores and ablation analysis for the segmentation based on different lexical features. Token-
based Accuracy figures are Strict when applied only to B-labeled paragraphs in the oracle, and Greedy when
also all the consistently aligned I-labeled paragraphs are considered as correct.
Basic + Simple + Lexical + Word
Editorial Lexical Contextual Embedding
Features Features Features Features
Precision 13.9% 44.4% 50.0% 52.1%
B tag Recall 34.3% 35.8% 34.3% 37.3%
F1 19.7% 39.7% 40.7% 43.5%
Precision 94.0% 94.9% 94.9% 95.1%
I tag Recall 82.9% 96.4% 97.3% 97.3%
F1 88.1% 95.7% 96.0% 96.2%
Token-based Strict 17.3% 33.4% 36.5% 40.5%
Accuracy Greedy 33.0% 70.7% 75.8% 77.0%
3 Experimental Evaluation and Discussion
In the current experimental stage, we applied ablation analysis to the set of Lexical Features as described in
the previous section. The lexical model is tested via the HMM SVM framework (implemented within KeLP,
[FCM+ 18]) on the annotated version of The Cherry Orchard, in its English translation: the Appendix reports a
short excerpt. The work includes 4 acts made of about 904 paragraphs, segmented into 67 units. The different
tokens in our labeled corpus L = OL are thus about 22,800. Every paragraph has been considered part of a
sequence of length k=5 that corresponds to the local input to the tagger. Every paragraph in the sequence is
represented via feature vectors and several lexical representations have been adopted:
• (Simple Lexical ) A bag-of-word feature vector including lemmas, bi-grams and POS tags occurring in the
target paragraph are represented.
• (Baseline) As a baseline, a set of simple heuristics from narrative features is used to simulate a blind
typographic approach. The synthetic vector encodes only a label with the guessed editorial role of the
paragraph. In this way, individual utterance and editorial notes are just kept separated.
• (Contextualized lexical ) Similar to the Lex feature vector, but extended with the vector of the preceding
paragraph, in order to contextualize the model.
• (Word Embeddings) A real-valued vector that corresponds to the sentence embedding of the target paragraph
is adopted.
Each training paragraph is labeled according to the IOB notation, i.e. as B or I. A macro n-fold cross validation
is applied with one fold per act, i.e n = 4. In one evaluation step, one act is removed from the dataset: training
on the three remaining acts is carried out by leaving a 10% of paragraphs as development data (i.e. for tuning
of the SVM parameters): the automatic tagging over the left-out act allows to measure and average the labeling
accuracy. Each time, the training set is randomly split 90/10 to derive a development set used to tune the SVM
parameters.
Measures of performance are class-based precision and recall, while accuracy is the percentage of paragraphs
that are correctly re-labeled with respect to the original IOB label. Micro-average across the four different 4
folds is applied. Notice that for the unbalanced presence of the I tag (i.e. 92.6% of the paragraphs), the simple
baseline model achieved a 93.5% of accuracy across all paragraphs. For this reason, in Table 1, we just report
precision, recall and F1 for the two separated classes. Moreover, we report the strict accuracy, just computing
the accuracy restricted to the B gold labeled paragraph. Notice that this class is defined by only 67 positive
examples in the training dataset. Finally, the accuracy measured on only the aligned B and I paragraph is
reported: it considers correct a paragraph labeled as “inner” by the system only when this does not violate any
boundary B in the oracle annotation. As Table 1 shows more complex lexical features brings more information
as they increase performance for each measure. Moreover, (last column in last line) the token based accuracy
suggests that the current model correctly annotate about 70% of the paragraphs of the work, thus representing
a large advantage against manual annotation.
Examples of mistaken segmentations are hereafter reported where the gold and automatic labels are shown
after the row number for the different paragraphs, respectively:
...
801 I I (Goes off )
802 B I YASHA remains, sitting beside the shrine.
803 I B Enter RANYEVSKAYA, GAYEV, and LOPAKHIN.
804 I I LOPAKHIN. It has to be settled once and for all - time won’t wait. Look, it’s a simple enough question.
Do you agree to lease out the land for summer cottages or not? Answer me one word: yes or no? Just one word!
...
901 I I YEPIKHODOV (Off, behind the door ). I’ll tell about you!
902 I I VARYA. Oh, coming back, are you? (Seizes the stick that FIRS left besides the door.) Come
on,then...Come on... Come on... I’ll show you... Are you coming? My word, you’re going to be for it...!
(Raises the stick threateningly.)
903 I B Enter LOPAKHIN.
904 B I LOPAKHIN. Thank you kindly.
905 I I VARYA (angrily and sarcastically). Sorry! My mistake.
...
According to the gold labels, the B-labeled paragraphs in lines 803 and 903 are all wrong, while 801,804
and 901,902,905 are correctly aligned I-labeled paragraphs: these latter are retained in the greedy version of
the Token-based Accuracy scores. Notice how the mistakes are mainly due to mismatches in the way editorial
material is used by the human annotators. In the first example, lines 802, 803, the beginning is annotated
to the sitting act of YASHA (line 802). In the second, the first speech of LOPAKHIN is used to start a new
segment (line 904). On the contrary, in both cases the system has focused on the entrance of the new character
to suggest the start (i.e. B labels in line 803 and 903).
These mild errors suggest how the generalization of the system at this current stage of development is already
acceptable in several cases. Accuracy rates are thus expected to grow when more complex features (for example
the narrative features that will better express the ontological information) will be adopted. This will be part of
future work.
References
[BCV+ 16] Emanuele Bastianelli, Danilo Croce, Andrea Vanzo, Roberto Basili, and Daniele Nardi. A discrimi-
native approach to grounded spoken language understanding in interactive robotics. In Proceedings
of the Twenty-Fifth International Joint Conference on Artificial Intelligence, IJCAI 2016, New York,
NY, USA, 9-15 July 2016, pages 2747–2753, 2016.
[CB11] Danilo Croce and Roberto Basili. Structured learning for semantic role labeling. In AI*IA 2011:
Artificial Intelligence Around Man and Beyond - XIIth International Conference of the Italian Asso-
ciation for Artificial Intelligence, Palermo, Italy, September 15-17, 2011. Proceedings, pages 238–249,
2011.
[Che17] Anton Chekhov. The Cherry Orchard. Plays, by Anton Tchekoff. 2d series, tr. with an introduction
by Julius West. New York, Scribner’s, 1917.
[Cho00] F. Y. Choi. Advances in domain independent linear text segmentation. In Proceedings of the 1st
NAACL Conference, pages 26–33. ACL, 2000.
[DLPar] Rossana Damiano, Vincenzo Lombardo, and Antonio Pizzo. The ontology of drama. Applied Ontol-
ogy, to appear.
[FCM+ 18] Simone Filice, Giuseppe Castellucci, Giovanni Da San Martino, Aless, ro Moschitti, Danilo Croce,
and Roberto Basili. Kelp: a kernel-based learning platform. Journal of Machine Learning Research,
18(191):1–5, 2018.
[Hea94] Marti A. Hearst. Multi-paragraph segmentation of expository text. In ACL, pages 9–16. Morgan
Kaufmann Publishers / ACL, 1994.
[KKM98] Min-Yen Kan, Judith L. Klavans, and Kathleen R. McKeown. Linear segmentation and segment
significance. In VLC@COLING/ACL, 1998.
[LDP18] Vincenzo Lombardo, Rossana Damiano, and Antonio Pizzo. Drammar: A comprehensive ontological
resource on drama. In ISWC 2018 - 17th Int. Semantic Web Conf., Monterey, CA, USA, October
8-12, 2018, Proceedings, Part II, pages 103–118, 2018.
[LPD16] Vincenzo Lombardo, Antonio Pizzo, and Rossana Damiano. Safeguarding and accessing drama as
intangible cultural heritage. JOCCH, 9(1):5:1–5:26, 2016.
[LSJ18] Jing Li, Aixin Sun, and Shafiq Joty. Segbot: A generic neural text segmentation model with pointer
network. In Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelli-
gence, IJCAI-18, pages 4166–4172. International Joint Conferences on Artificial Intelligence Organi-
zation, 7 2018.
[MCCD13] Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. Efficient Estimation of Word Represen-
tations in Vector Space. CoRR, abs/1301.3781, 2013.
[Rey94] Jeffrey C. Reynar. An automatic method of finding topic boundaries. In ACL, pages 331–333.
Morgan Kaufmann Publishers / ACL, 1994.
[SDDK11] Fei Song, William M. Darling, Adnan Duric, and Fred W. Kroon. An iterative approach to text
segmentation. In ECIR, volume 6611 of Lecture Notes in Computer Science, pages 629–640. Springer,
2011.
[STC04] John Shawe-Taylor and Nello Cristianini. Kernel Methods for Pattern Analysis. Cambridge Univer-
sity Press, 2004.
[SZR07] N. K. Speer, J. M. Zacks, and J. R. Reynolds. Human brain activity time-locked to narrative event
boundaries. Psychological Science, 18(5):449–455, 2007.
[TJHA05] Ioannis Tsochantaridis, Thorsten Joachims, Thomas Hofmann, and Yasemin Altun. Large margin
methods for structured and interdependent output variables. J. Machine Learning Reserach., 6,
2005.
[Yaa99] Yaari Yaakov. Segmentation of expository text by hierarchical agglomerative clustering. In Recent
Advances in NLP (RANLP’97). ACL, 1999.
[YXX+ 16] J. Yu, X. Xiao, L. Xie, E. S. Chng, and H. Li. A dnn-hmm approach to story segmentation.
INTERSPEECH, pages 1527–1531, 2016.
[ZS07] J. M. Zacks and K. M. Swallow. Event segmentation. Current Directions in Psychological Science,
16(2):80–84, 2007.
[ZSSM10] J. M. Zacks, N. K. Speer, K. M. Swallow, and C. J. Maley. The brain’s cutting-room floor: Segmen-
tation of narrative cinema. Frontiers in human neuroscience, 4, 2010.
[ZT01] J. M. Zacks and B. Tversky. Event structure in perception and conception. Psychological bulletin,
127(1):3, 2001.
Appendix: a segmentation example: unit 4-6
From The Cherry Orchard by Anton Checkov
0034 YEPIKHODOV. I’ll go, then. (Stumbles against the table, which falls over.) There you are... (As if exulting in it.)
You see what I’m up against! I mean, it’s simply amazing! (Goes out.)
UNIT ID: 0004, UNIT NAME: Dunyasha struts around, SPAN:35,37
0035 DUNYASHA. To tell you the truth, he’s proposed to me.
0036 LOPAKHIN. Ah!
0037 DUNYASHA. I don’t know what to say... He’s all right, he doesn’t give any trouble, it’s just sometimes when
he starts to talk you can’t understand a word of it. It’s very nice, and he puts a lot of feeling into it, only you can’t
understand it. I quite like him in a way, even. He’s madly in love with me. He’s the kind of person who never has any
luck. Every day something happens. They tease him in our part of the house - they call him Disasters by the Dozen...
UNIT ID: 0005, UNIT NAME: Lopakhin and Dunyasha welcome the masters, SPAN:38,42
0038 LOPAKHIN((listens). I think they’re coming.
0039 DUNYASHA. They’re coming! What’s the matter with me? I’ve gone all cold.
0040 LOPAKHIN. They are indeed coming. Let’s go and meet them. Will she recognize me? Five years we haven’t seen
each other.
0041 DUNYASHA (in agitation). I’ll faint this very minute... I will, I’ll faint clean away!
0042 Two carriages can be heard coming up to the house. LOPAKHIN and DUNYASHA harry out. The stage is empty.
UNIT ID: 0006, UNIT NAME: The owners settle down, SPAN: 43, ...
...