Levels of text analysis Flashcards
Discourse vs. Words vs. Sentences
- Discourse – framing a situation; a path from a beginner to an expert (in the context of the bank quiz); learning finance is like learning language, etc.
- Words – metaphors, choice of words, frequency, collocation, etc.
- Sentence – a triplet of agent, action, and object; the argument; complexity, etc.
2 approaches of text analysis
- Textual analysis
2. Text reception analysis
Textual analysis
3 levels
- Lexical level – the level of lexemes (words); word frequency
- Level of sentence – the level of linguistic acts; meaning; argument
- Supra-sentential level – the level of discourse; sequence of sentences
Text reception analysis
- The readers’ cooperation with texts (author’s textual strategies and reader’s reading strategies)
- How readers interact with text
- Sometimes what’s written in the smallest letters is most important
- Used to inhibit will to read it; done on purpose; not about the content but about accessibility
The role of the reader
3 main roles
- The reader’s role is crucial for any text analysis because:
- Readers assign textuality to an object
- Readers construct intertextual, transtextual, and contextual relations
- Readers establish the meaning of a text
- The text itself can’t assign any meaning, analysis, relationships to itself
- Authors can prepare material in a way that makes connections obvious or in ways that provoke thoughts
Operative vs. Cooperative reading
- Distinguished by U.Eco
Reading requires certain competencies from the reader:
- Reader as an Operator
* Only a linguistic competence (grammar, lexicon) is necessary to actualize a lexemic surface of the text - Reader as a Co-operator
* Linguistic, conferential and encyclopedic competencies are needed to actualize a deeper meaning of the text
- Linguistic - you need to know english
- Coreferential - need to understand lack of sense from these repeating words
- Encyclopedic - need to know reference to “to be or not to be” and that it refers to life vs death
- Qualitative analysis - co-operator
- Quantitative analysis - operator
What’s the model reader?
A model reader actualizes the meaning of everything that the textual strategy (of the author) intends to say
- Model reader can read the text as the author intended it
- Should the analyst be a model reader for a text he/she analyzes?
- A good author give signs and hints to be understood as they intended
- Author needs to write to be understood properly
Tony Bennet’s strategies of co-operative reading
- Some texts (metatext) organize reading of other texts.
- They help the reader (or prevent him from) becoming a model reader for other texts
Examples
- Micro-level – prefaces, preambles, introductions, notes, commentaries, manifests, caricatures, examples, etc.
- Macro-level – Marxism, psychoanalysis, hermeneutics of suspicion, conspiracy
- Once you read a text, you will always have that knowledge behind how you analyze future texts
- Hermeneutics of suspicion - text was created to sell me something or convince me of something; readers then read with suspicion
- Conspiracy - conspiracy thinking makes you a model reader because
Stuart Hall’s strategies of co-operative reading
- Reading strategies are relative to the social context of the reader’s position
- Developed for analyzing TV
Examples
- Dominant reading - the reader accepts the socially preferred mode of reading
- Negotiated reading - the reader accepts in part the preferred mode but in part reads alternatively; Some criticism
- Oppositional reading - the reader reads a text in a contrary position toward it
Lexical level of analysis
What is a word?
* A word (lexeme) - elementary unit of a text that generates meaning
* A “table”
= a piece of furniture with a flat top and one or more legs, providing a level surface for eating, writing, or working at;
= a set of facts or figures systematically displayed, especially in columns.
* Even a single letter can generate meaning like “B” being the label referring to an item, and so they function as words
Two types of words
- Content words – having meaning (table, to give up, clever); mostly used in textual analysis
- Function words – Express grammatical relationships with other words (like, a, because); rarely used in textual analysis
Lemma
- The canonical form for a family of words
- Teach stands for teaches, taught, teaching
- Proxy for analysis on the level of words
- Lemmas are important for inflected languages as Slavic languages
Keyword
- A natural representation of a semantic field
- stand for some field of meaning
- “Nano” stands for nanotechnology, nanospider, nanofiber, and other related words
- Global (for things that concern the whole earth), bio/organic, poverty, gender…
- Represent some thematic or semantic fields
Code
- Most artificial replacement of words
- Constructed index for a semantic field
- Example “school” stands for all words linked to education
- School, teacher, pupil, etc…
- Used for quantitative analysis
What can we do with words (keywords, lemmas, etc.)?
- We can count them, observe their frequency in our samples, and relate them to the (social) context
- We can observe their occurrence in relation to their co-text (collocations, co-occurrence, and KWIC (Key Words In Context))
Co-location
Some words occur together more often than randomly
- “After 9/11”, “undocumented immigrants”, “latest news”
Co-occurrence
- Not just that two words are close together
- Counting how many times two or more words occur together in a determined span of a text (article, paragraph, sentence, webpage)
- Not closely related but found close together in some span of text
Keyword-in-context (KWIC)
A list of co-texts of a word in the text (corpus). Helps to identify:
- the frequent/typical use of the word
- non-frequent use of the word
- positive/negative tone
Levels of a sentence
This is the primary difference: words pertain to language while sentences pertain to authors or speakers
We can use a dictionary for words, but there’s no dictionary for sentences
* We don’t make words but we do make sentences
* Sentences are results of our communicative intention
* They’re OUR sentences
If we don’t understand, we have to ask the person what they meant
Sociological Text Analysis uses a pragmatic mode of analysis rather than syntactic or semantic - it focuses on what people do by using sentences, (like negotiating, agreeing, convincing, attacking, harming, etc)
Semiotics has 3 branches:
- Semantics (meanings)
- Syntactics (the combination of and relationship of symbols)
- Use of language in society
We can analyze sentences 3 ways:
- Semantically (what do they mean?)
- Syntactically (how are they constructed?)
- Pragmatically (how do people use sentences to do something?) - used for T.A.
2 models of communication
- Encoding – decoding
2. Inferring
Encoding – decoding
The sender wants to communicate, they encode in a message (like language), and the receiver decodes (understands)
- “How are you?”
- The classical model of communication
- Not the same as Stuart Hall’s chapter
Inferring
The decoder also has to infer the meaning in addition to the language (mostly from social context - tone, body language, etc.)
- “I’m fine”
- In sociology, the communication model doesn’t work
- We don’t know the code in which it was encoded
- The decoder could interpret “I’m fine” in many ways
- It’s impossible to know from just language
Supra-sentential level
Meaning “the level above the sentence”
Highest level of analysis
* Focus on the context in which sentences occur or look for macrostructures (storytelling, is there a story? Just description or plot? What’s the plot? What is the point? What’s interesting about it?)
Texts have properties that can’t be explained
- From the properties of sentences
- In linguistic terms
What makes words distinct from sentences? (The answer is not that sentences consist of words.)
Words pertain to language while sentences pertain to authors or speakers. We don’t make words, and so we can look them up in the dictionary if we don’t understand. But we do make sentences, so we have to ask the speaker for their communicative intention if we don’t understand.
Comparing the number of occurrences of the keyword “globalisation” in newspapers with the amount of international trade for the same period of time is an example of word co-occurrence study. True or false?
False
Why encoding/decoding model of communication is insufficient for the analysis of discourse (the use of language in society)?
The encoding and decoding model of communication doesn’t work for social science because we don’t know the code in which messages were encoded. For example, a decoder could interpret “I’m fine” in many ways and it’s impossible to know from just language. Rather, the decoder also has to infer the meaning in addition to the language (mostly from social context - like tone, body language, etc.)