Session 3: Hands-on #1

Basic Concordancing

Masaki EGUCHI, Ph.D.

Housekeeping

Session overview

  1. Corpus lab assignment overview
  2. Corpus Research Cycle (Cont’d)
    1. Operationalization
    2. Testing
    3. Interpretation

Corpus lab 1

Goals and Overview

This assignment aims to help you practice the following skills:

  • Articulating linguistic research questions for corpus-driven research
  • Explaining corpus choices and analysis approach for basic corpus search
  • Interpreting and describing the basic concordance search results

Goals and overview

  • Task 1: Constructing corpus-search research question and hypothesis
  • Task 2: Methods
  • Task 3: Result

Task 1: Research question and hypothesis

In this task, you are asked to articulate your research question and hypothesis for your first corpus assignment.

Research question:

  • Research questions should be answerable.
  • Write one (or two related) research questions you would like to answer through this assignment.

Task 1: Research question and hypothesis

Hypothesis:

  • Once you decided on research question, state your hypothesis.
  • Write a short paragraph stating your hypothesis and why you think your research hypotheses may be true.

Task 2: Methods

In this task, you are asked to describe the methods choice of your corpus search. This should include justifications for the corpora used, type of corpus search used.

Corpora

  • Which corpora or sub-section of a corpus would you conduct a search on? Why? Justify your choice in a paragraph.

Task 2: Methods (cont’d)

Search Methods and Plan

  • What corpus search methods would you choose and why? In this assignment, search methods mainly include methods available through English-Corpora.org, including LIST, Chart, Collocates, Compare and KWIC.
  • In what way are you planning to conduct the search and what kind of information are you expecting from it?

Task 3: Result

In this part of the assignment, you are asked to describe the search results and provide interpretations on the findings.

  • Corpus search results: Provide some numbers or discourse samples based on your corpus search. This can be frequency list, table with frequency counts, or a copy of KWIC results.

  • Interpretation: Write a paragraph, providing some interpretations of your findings.

Submission

Success Criteria

Your submission …

Recap: The Scientific Research Cycle

Remember from Session 2:

  1. Verbalize your question
  2. Formulate research questions
  3. Formulate hypothesis
  4. Choose appropriate corpus
  5. → Operationalize
  6. → Test & Evaluate

What does “operationalize” mean?

Operationalization

“an explicit and unambiguous description of a set of operations that are performed to identify and measure that construct.” (Stefanowitsch, 2020, p. 77)

  • Defining exactly WHAT to search
  • Deciding HOW to search it
  • Being explicit about your choices

Measure of height

  • How to operationalize “height”?
    • Meter? Feet?

Result of the search with run

  • Yes, you now get the frequency result (= 264014).
  • You might get slightly different results.

Search - run

Let’s look at the context

  • Wait a minute… where is runs, ran, and running? run in context

Thinking about the operationalization

  • Now we recognized that the word run was just one of the possible forms of the headword run.

  • If we were thinking to retrieve all the occurence of the word “run” then we were imprecise.

  • What should we do?

Lemma

  • Lemma: A headword and all the inflected form derived from it.

    RUN:

    • run, ran, runs, running… etc.

Refining our operationalization

  • Concept: All forms of “run”
  • Better operationalization: Search for LEMMA

LEMMA search

Searching lemma - run

LEMMA search result - run

Search result

Specifying Part Of Speech (POS)

You can use POS tag like the following .

Category Simple tag Symbol Example
Common noun NOUN N run_N
Proper Nouns NAME NP Sendai_N
All nouns NOUN+ N+ sun_N+, Sendai_N+
Lexical verbs VERB V run_V
All verbs VERB+ V run_V+, do_V+

Specifying Part Of Speech (POS)

Category Simple tag Symbol Example
Adjectives ADJ J simple_J
Adverbs ADV R clear_R

See this page for more information of POS in English-Corpora.org

the word run used as noun

  • Search occurrences of run that are used as nouns.

the word run used as noun

Search: run_N

run - as noun

Your turn: Operationalize these concepts

  • The word “chair” as a verb only

Task B - Basic KWIC

Concordances (or KWIC)

  • Now we know frequency of X in COCA, but what about context?

KWIC = Key Word In Context

KWIC will give us insights into how each word is used in context.

KWIC view

  1. Go back to SEARCH window.

KWIC search

KWIC view

  1. Click on the + button in the search menu and enter your search word.

KWIC search

KWIC view

  1. The result is displayed. Default sort : R1 > R2 > R3

KWIC search

Sorting KWIC window

  • Now let’s sort the results according to the position in context.

KWIC search

Sorting words

You can sort the words.

KWIC search

Sorting to find patterns

Sorting helps identify: - Common phrases - Grammatical patterns
- Semantic preferences

Let’s Try: KWIC

  • Choose a word that you want to see the context for.

  • Search the word with KWIC.

  • Sort the word in the following way.

    • Default: R1 > R2 > R3
    • Custom 1: L1 > L2 > L3
    • Custom 2: L1 > R1 > R2

How to get Custom 1 and 2

KWIC search

Practice: Finding patterns

Search “prove” (lemma) and sort by: 1. R1 (what follows “prove”?) 2. L1 (what precedes “prove”?)

What patterns do you notice?

Task C

Research question: How does frequency of lol change across time?

Chart function

CHART shows frequency across:

  • Genres

  • Time periods

  • Text types

  • This allows you to return frequency by sections of of COCA (= conditional frequency).

Search lol using CHART

  • Go to Search Tab, select, CHART and enter lol

Search lol using CHART

  • You will get the following:
  • What do you notice here?

Search - lol

Interpreting the distribution

What does this tell us about “lol”?

  • Genre preferences?

  • Formality levels?

  • Change over time?

  • Describe the frequency pattern looking at PER MIL row.

Summary

CHART function can be used to get frequency across:

  • Genres
  • Time periods
  • Text types

Task D

Research question: What is most frequent word that ends with derivational mopheme -ness?

Regular expressions

Regular expression (正規表現) allows you to search corpus through “pattern matching”.

  • Have you ever used (*) asterisk in your web search?
  • This is one of the regular expression (= wild card)

Using regular expression to get words with specific morpheme

Any idea for operationalizing derivational morphemes?

  • Operationalization: *ness

Results of *ness

Search results *ness

Using wild card for lexicalized pattern

  • Let’s go back to List search.

  • Enter “a * of the”

  • What result do you expect with this search?

  • Don’t turn to the next page YET!!!

Result with a * of the

  • a * of the

This helps find:

  • Fixed phrases with variable slots
  • Grammatical frames
  • Idiomatic expressions

More search methods

  • You can check how to use other search methods in English-Corpora.org here

More complex searches

Which is more frequent?

  • exciting or excited?

  • RQ: Is there any reletitive phrase with three adjectives?

ADJ x 3

three adjectives

  • Note: Sometimes the search results will contain parsing errors. (We will come back to this topic on Day 4.)

Task E - Collocation (Skip if no time)

Collocation search

  1. Select collocates

collocation

Collocation search

  1. Enter search word

entering words

Collocation search

  • Optional enter the word to look for

enter collocates

Collocation search

  1. Specify window
  • How distant do you search for the collocates

window

Search Methods Summary

Method Purpose Example Key Features
Simple Word Find exact word forms run → finds only “run” - Case sensitive
- Single form only
LEMMA Find all forms of a word RUN → finds “run, runs, ran, running” - Use CAPITAL letters
- Includes all inflections
POS Tag Find words by grammatical category run_N (noun)
run_V (verb)
- Disambiguates word classes
- Use underscore + tag

Search Methods Summary

Method Purpose Example Key Features
KWIC View words in context Any search term - Sort by L1, R1, etc.
- Find patterns in usage
CHART Track frequency across categories lol across time/genres - Shows distribution
- Genre/time comparisons
Wildcards Pattern matching *ness → “happiness”
a * of the → “a lot of the”
- * = any characters
- Find lexical patterns
Collocates Find co-occurring words Words near target - Specify window size
- Statistical associations

Bringing It All Together: Corpus Lab 1

Your Assignment Structure

  1. Research Question & Hypothesis (Sessions 1-2)
  2. Methods (Session 2-3)
    • Choose corpus
    • Operationalize search methods and terms
  3. Results (Session 3)
    • Conduct searches
    • Interpret findings

Research question:

  • Should be answerable with corpus data
  • Focus on distribution/frequency/context

Hypothesis:

  • Make a prediction
  • Explain your reasoning using the template we have used

Task 2: Methods

Corpus Selection

  • Which corpus/corpora?
  • Why did you choose it/them?
  • What does it allow you to test?

Search Plan

  • How will you operationalize your concepts?
  • LIST, CHART, KWIC, or combination?
  • Any special search syntax needed?

Task 3: Results

Present your findings:

  • Frequency tables
  • KWIC samples
  • Distribution charts

Interpret the patterns:

  • What do the numbers mean?
  • What do the contexts reveal?
  • Does this support your hypothesis?

Submission

Success Criteria

Your submission …

Start thinking now!

What patterns in English interest you?

Consider:

  • Words that seem to be changing
  • Regional differences
  • Formal vs. informal patterns
  • New expressions
  • Grammar variations

Today’s achievement

  • Operationalize linguistic concepts for corpus search
  • Use different search types (word, LEMMA, POS)
  • Analyze patterns through KWIC
  • Evaluate distributions with CHART
  • Apply the full research cycle