A Discourse Analysis

Determining the Storyboard Assets of Indonesian Environmental Public Service

Animation Scripts

Fathimah Salma Zahirah

Department of Linguistics,Universitas Pendidikan Indonesia, Bandung, Indonesia

s.zahirah90@gmail.com

Keywords: Script, Storyboard, Assets, Content Words, Syntax, Discourse Analysis, Mood.

Abstract: An animation can be categorized as a good media if it is created through an elaborative storyboard-making

procedure. The most important thing in the storyboard production is making the items (assets) that will be

animated based on the script. In other words, the script must contain some content words (e.g. bed, king, table,

etc.) that will be the basics on shaping the assets. Therefore, the advanced scanning process of the content

words needs to be taken into account in order to make the storyboard-making process quicker. Based on this

background, this research paper elaborates a syntactic discourse analysis, in which a mood classification as a

systemic functional grammar tool also plays a role in order to find the content words from the script. After

examining the script in that manner, the content words were interpreted as visual products. This research paper

employed ten 2017 Indonesian environmental public service animation scripts as the data. The further attempt

will lead into assessing the possibility of making a program system for creating assets visualization.

1 INTRODUCTION

Halliday (Paltridge, 2006) argues that a written

discourse tends to be more lexically dense than a

spoken discourse. Lexical density refers to the ratio

of content words to grammatical, or function words,

within a clause. Content words include nouns and

verbs while grammatical words include items such as

prepositions, pronouns, and articles. There is also a

high level of nominalization in written texts where

actions and events are presented as nouns rather than

as verbs. Written texts also typically include longer

noun groups than spoken texts.

In that sense, a public service script can be

categorized into a persuasive spoken discourse. This

kind of discourse tends to have statement sentences.

Even though it is written, the sentence structure on the

script is more into that of the spoken one. Aside from

being complex, there are more simple sentences

found on the script. It leads to an assumption that

some sentences may lack of subjects because the

nature of the script tends to employ a lot of imperative

sentences; since it is a spoken discourse.

Regarding the presence of subjects, Gerot and

Wignell (1994) state that the mood element –Subject

+ Finite– carries the burden of the clause as an

interactive event. Thus, it is plausible that the

sentences contain no mood. The mood system on the

animation script is important for determining which

scene that should be highlighted. Moreover, what

kind of effects and features are suitable for the

animation assets are also contextualized through the

mood. This consideration plays crucial matter

whether the content of the script should have a

concrete visual form or stay as texts only. Hence, not

only the syntactic feature of the sentences should be

taken into account but also the functional one.

In relation to the script representation, Hart (2008)

states that the storyboard is the premiere

preproduction, pre-visualization tool designed to give

a frame-by-frame, shot-by-shot series of sequential

drawings adapted from the script. They are concept

drawings that illuminate and augment the script

narrative and enable the entire production team to

organize all the complicated action required by the

script before the actual animating is done to create the

correct look for the finished motion.

Based on that background, this paper will discuss

the possibility of creating a program system that can

denote the content words automatically. Mason and

Charniak (2012) created a system for connecting

corpus to images based on Natural Language

Processing Systems. The system can be more precise

Zahirah, F.

A Discourse Analysis - Determining the Storyboard Assets of Indonesian Environmental Public Service Animation Scripts.

DOI: 10.5220/0007163501450148

In Proceedings of the Tenth Conference on Applied Linguistics and the Second English Language Teaching and Technology Conference in collaboration with the First International Conference

on Language, Literature, Culture, and Education (CONAPLIN and ICOLLITE 2017) - Literacy, Culture, and Technology in Language Pedagogy and Use, pages 145-148

ISBN: 978-989-758-332-2

145

if it is engaged with Moeljadi’s (2017) Lexical

Functional Grammar Treebank (JATI). However,

none of both focuses on a certain type of text and

image.

2 METHODOLOGY

There are ten Indonesian scripts taken as the data. The

scripts were first analysed based on the types of the

sentences by using the mood framework. The content

words found were classified into denotative and

connotative interpretation. The verbs with affixes

were interpreted based on their lexemes. Thus, the

sentences that contain a lot of function words,

especially the one with interpersonal sense, would be

regarded as a merely text when they are realized

through the storyboard–did not need to be visualized.

2.1 Types of the Sentences

The nature of a persuasive text is to give information

and make the readers aware of their current actions

regarding the text. This type of text uses a lot of

statements to justify the purpose. The result shows

that seven scripts use statement sentences

productively (52% approximately). Meanwhile, the

other three scripts tend to be more commanding (50%

approximately).

2.2 Mood Analysis

The mood analysis, as one of the structural

framework, gives insight to the importance of

interpreting a word in relation to its context. Chaer

(2006) states that a language is a connected system,

in which each of its elements –sounds, words, etc.–

does not have any validity except within the

structured relation. However, the interpretation of the

word itself can be seen from dyadic perspective, as

proposed by Saussure (Hoed, 2014): “the relation

between sign can be syntagmatic and paradigmatic”.

Hence, the result of the mood analysis will be

connected to the interpretation of the content words

later on.

The data show that there are 31.48% of mood

elements found in the scripts. The scripts which have

more statement sentences are likely to have more

mood aspects. In contrast, the other scripts which

contain more command sentences are likely to have

more residues. Although there is such diversity, the

amounts of content words (noun, noun phrase, verb,

and verb phrase) are randomly varied within each

script.

2.2.1 Mood

In general, mood consists of a subject and finite. The

former has the higher authority in creating mood for

a static visualization of a noun. Meanwhile, the latter

is more applicable for determining certain temporal

nuances when it comes to the animated version of a

noun. If the subject is a content word (whether it is

followed by finite or not), it will be necessary to

contrast it from the other elements when it is

represented in a storyboard.

Based on the data, the main mood in the scripts is

building the awareness of maintaining our

environment. The elements that should be highlighted

visually are mostly ‘sampah (garbages)’ which occur

48 times.

2.2.2 Residue

Residues of the sentences also play important role in

interpreting the script. The residue contents help the

storyboard artist decide which elements that should

be put as the background or the visual complements.

For example, in “Sampah yang dibakar malah

melepaskan zat-zat yang berbahaya!“, the residue is

“zat-zat yang berbahaya (dangerous substances)”.

This means, the dangerous substances will have a

minor appearance.

Furthermore, the data also indicate that the more

command sentences a script has, the less likely the

intended information will be well-received.

2.2.3 Interpersonal Content

The interpersonal content may not be concretely

visualized. However, it creates a friendly tone for a

public service animation. This content is likely to be

represented typographically or via voice over. If the

script has a mascot, the interpersonal content can also

be appeared in one scene with it.

The data show that besides employing question

sentences to build interpersonal relationship with the

audiences, there are also some pragmatic fillers such

as ‘nah…(well)’, ‘jadi…(so)’, and ‘dan…(and)’.

Moreover, some of the interpersonal contents found

in the data are put separately from the previous

sentence. They tend to indicate transitions.

2.2.4 Content Words

There are 19.34% of content words found in the

scripts, from 736 total words. The amount of content

words is not necessarily positive with the amount of

the total words. It means, the script can have a lot of

CONAPLIN and ICOLLITE 2017 - Tenth Conference on Applied Linguistics and the Second English Language Teaching and Technology

Conference in collaboration with the First International Conference on Language, Literature, Culture, and Education

146

messages but they are represented through the same

assets.

Moreover, even though noun phrase and verb

phrase are syntactically regarded as content words,

there might be a problem to represent the verb phrase

in a storyboard. Especially, the one contains an action

process. Thus, most of the storyboard assets are likely

derived from noun phrases, even only nouns as

lexemes.

2.3 Determining Content Words

Visualization

The possibility of making image annotation based on

a text becomes a promising project for those who

work in an animation field. The previous studies from

Mason and Charniak (2012) and Moeljadi (2017)

already show that the idea of making contextual

image annotation is plausible. However, there are

some considerations that need to be assessed.

2.3.1 Scheme

The most important thing is one should decide which

system can perform the image annotation faster. It is

likely that a system of text scanning should employ

texts as the data bank (corpus). Moreover, the system

should also have the main library of lexical

definitions as to detect the lexeme of a morphemic

word. The crucial matter will arise when the system

should relate the targeted word with its context. The

following discussion will elaborate more on how to

scan content words as efficient as possible. The

scheme of the system can be seen on Figure 1.

Figure 1: The System Scheme.

2.3.2 Text as the Input

The system should be able to detect content words

from the script that will be represented as assets on

the storyboard. If there is any Indonesian slang term,

the system should also be able to recognize it. Since

the system will only focus on words, there should be

a special treatment for numbers. It is necessary if the

numbers are represented as assets or they function as

content word adjectives (e.g. in “Denda mulai dari Rp

250.000,00”; not only ‘denda (fine)’ but also the

amount of the money is considered as a content

word).

2.3.3 Content Words Identification

The identification of Indonesian content words will

be easy if the lexeme is already stored in the metadata

of the system. However, the problem will occur if the

content words have complementary elements which

can only be tracked based on the context. Thus, the

best option for the content words fast scanning

system, in relation to a storyboard, is to determine the

assets on the lexemes. Even though it is possible to

create the contextual interpretation, there will be an

obstruction of the individual creativity.

Relying on that consideration, the safest option is

denoting the noun (without the verb) through the most

common social representation. Nevertheless, such

system still will be helpful because it decreases the

time consumption when scanning the scripts

manually for a production purpose. The content

words identification process can be seen on Figure 2

Figure 2: The Content Words Identification Process.

2.3.4 Stemming

The morpho-syntactic analysis will be the lead for

shortening the morphemic content word into its

lexeme. The reasonable process for that purpose is

word stemming. The morpheme(s) attached to the

word will be eliminated in order to find the lexeme

form. The most suitable model for this operation is

A Discourse Analysis - Determining the Storyboard Assets of Indonesian Environmental Public Service Animation Scripts

147

Halle’s morphological organization in Ba’dulu and

Herman (2005) as shown in Figure 3.

Figure 3: Halle’s Morphological Organization.

In order to avoid an overstemming process, the

dictionary put in the metadata should define the

words that may seem morphemic but indeed they are

lexemes. For example, the word ‘penyu (tortoise)’ is

not derived from ‘peny-u’. Moreover, a reverse logic

of a morphemic word creation should be taken into

account for eliminating allomorph. Therefore, the

system will give more credible lexeme forms. For

example, the word ‘menyapu’ has ‘sapu’ as the

lexeme, not ’nyapu’.

2.3.5 Treating Repeated Words

The repeated words found in the scripts should not be

presented more than one time from the text extraction.

Thus, there should be another ‘dictionary’ that can

add the scanned lexemes and make the system skip

the same lexemes which occur twice or more later on

(the process can be seen on Figure 2).

The system presentation sample can be seen on

Table 1.

Table 1: Content Words System Presentation Sample.

Extracted Content

Words

Visual Representation

Jam

3 CONCLUSIONS

It can be concluded that the mood analysis can help

interpreting the visual aspect of content words. Even

though it might be difficult to connect the functional

aspect to the syntactic aspect of a word in a program

system, it is still possible to employ those notions

together for determining the content words

visualization.

4 RECOMMENDATION

The system discussed in this paper is still a model. It

is recommended for the next researcher to develop the

prototype which is applicable for the storyboard-

making production. Other visual variables such as the

tone of reference and the colour schemes guidelines

for the storyboard also need to be considered.

ACKNOWLEDGEMENTS

The author would like to thank Bonbin Studio for

giving a privilege to analyse the environmental public

service animation scripts. Many regards are also sent

to Syarifuddin and Aria Dhanang for the discussion.

REFERENCES

Chaer, A., 2006. Tata Bahasa Praktis Bahasa Indonesia,

RINEKA CIPTA. Jakarta, 2

edition.

Ba’dulu, A.M. and Herman, 2005. Morfosintaksis,

RINEKA CIPTA. Jakarta, 1

edition.

Gerot, L., Wignell, P., 1994. Making Sense of Functional

Grammar: an Introductory Workbook, Gerd Stabler.

New South Wales.

Hart, J., 2008. The Art of the Storyboard: A filmmaker’s

Introduction, Elsevier. United States of America, 2

edition.

Hoed, B.H., 2014. Semiotik & Dinamika Sosial Budaya,

Komunitas Bambu. Depok, 3

edition.

Mason, R., Charniak, E., 2012. Apples to Oranges:

Evaluating Image Annotations from Natural Language

Processing Systems. In 2012 Conference of the North

American Chapter of the Association for

Computational Linguistics: Human Language

Technologies. Association for Computational

Linguistics.

Moeljadi, D., 2017. Building JATI: A Treebank for

Indonesian. Proceedings of Atma Jaya Conference on

Corpus Studies. Atma Jaya.

Paltridge, B., 2006. Dicourse Analysis: an Introduction,

Continuum. New York, 1

edition.

CONAPLIN and ICOLLITE 2017 - Tenth Conference on Applied Linguistics and the Second English Language Teaching and Technology

Conference in collaboration with the First International Conference on Language, Literature, Culture, and Education

148