Find every occurrence in its original context
Concordance searches the spoken words in your corpus and gives each occurrence its own result row. Every row preserves the surrounding language, speaker, corpus text, and transcript line so you can interpret the match and return to its source.
One occurrence becomes one traceable piece of evidence
An occurrence is one place where the requested word, phrase, or passage appears. If walk occurs 18 times, Concordance produces 18 rows. It does not reduce those rows to a single vocabulary entry.
Search one word, a sequence of adjacent words, or the passage between two literal markers.
Read what comes before and after each occurrence rather than treating the word as an isolated count.
Identify who produced the occurrence, which corpus text contains it, and where it begins in the transcript.
Choose the search scope with Active selection
The search scope is the part of the corpus included in the next search. Exempla uses the Active selection shown above the left panel tabs. The selection remains active when you move between Concordance, Word List, Sequences, and Corpus Statistics.
This is the default. Every imported corpus text is searched.
Select individual items in the Corpus tab when the question concerns particular interviews or documents.
Select codes in the Speakers tab when the question concerns particular people across all texts in which they appear.
Select one or more groups you created from corpus texts. Exempla searches the combined members of those groups.
Select one or more groups you created from speaker codes. Exempla searches utterances attributed to their members.
Click one item to select it. To select a continuous range, click the first item and Shift-click the last. To combine separate items, use Command-click on macOS or Ctrl-click on Windows. The Active selection summary confirms exactly what is included.
Use the small × in Active selection. This clears only the scope restriction. It does not delete a text, speaker, or group.
After the search, the status row begins with Searched texts. Open it to review the texts used for that completed search.
Run a Concordance search step by step
Choose Concordance in the left navigation. This opens the tool without starting a search or changing Active selection.
Leave Entire corpus active or select the texts, speakers, text groups, or speaker groups that answer your research question.
Click the search field and type the word, phrase, or pair of markers. A query is the text you are asking Exempla to find.
The Match menu changes automatically according to whether the query contains one token or several words.
Decide whether letter case, parenthetical comments, or joined apostrophe forms matter for this search.
Choose Search or press Return while the query field is active. Search remains unavailable until a corpus, valid scope, and valid query exist.
Read Searched texts, Tokens found, Time, and Order before interpreting the table.
Search one token or an exact sequence of tokens
A token is one countable word unit recognized by Exempla. Letters and numbers form tokens. An internal hyphen or period can remain part of a token, as in mother-in-law, U.S.A, or 3.14. Apostrophes follow the Join at apostrophes setting explained below.
Single-token queries
When the query contains one token, the Match menu offers the following word choices. Use the controls to compare them.
Exact word: Returns only the complete token “walk”. It does not return “walked”, “walking”, or “walkway”.
Returns only a complete token equal to the query. The default search ignores uppercase and lowercase differences.
Returns the exact query and longer tokens beginning with it. Searching walk can return walk, walked, and walking.
Returns tokens ending with the entered characters. Searching ing can return walking, growing, and meeting.
Returns tokens containing the entered characters anywhere. Searching alk can return walk, walking, and talked.
Queries containing spaces
As soon as the query contains whitespace, the Match menu changes to Exact phrase, Partial last word, and Between markers. Exempla does not require you to switch manually from a word search to a phrase search.
grew upFinds adjacent tokens in that order, such as “I grew up here.” Punctuation or a visual line break may separate the words, but an additional word may not.
New YorRequires every earlier token to match exactly, then accepts a final token beginning with the typed fragment. It can find New York and New Yorker.
- A phrase may continue across a line break introduced by the TXT file or Word wrapping.
- A phrase cannot jump over an intervening spoken word.
- A phrase cannot join the end of one speaker’s turn to the beginning of another speaker’s turn.
- Speaker codes are metadata and never become ordinary phrase tokens.
- Parenthetical comments are excluded unless Include comments is active.
Use Between markers for bounded passages
A marker can be any non-empty sequence of characters that appears in the transcript: a word, phrase, punctuation, or an annotation such as (en) or (fr). Between markers finds each start marker and the following end marker, then returns the passage between them. The markers themselves are not included in the matched Token value. Their exact characters and spacing matter; matching ignores letter case unless Case-sensitive is active.
This is useful for consistently tagged code-switching. If (en) marks the beginning of English speech and (fr) marks its return to French, searching (en) (fr) isolates the English stretches for inspection, counting as Concordance matches, and export. Parenthetical tags work here even when Include comments is off, because Between markers searches the literal source text.
Illustration
Separate two markers with ordinary whitespace. Example: (en) (fr).
Separate them with an ellipsis. Example: START MARKER … END MARKER. Three periods also work.
Use Up to to exclude passages containing more than the chosen number of countable words. Speaker codes and parenthetical comments do not count toward this limit. Any length applies no limit.
If a final start marker has no later end marker, Exempla returns the remaining text and identifies it as open-ended in the Token tooltip.
- Markers are matched literally. Their punctuation and spacing matter.
- Letter case is ignored unless Case-sensitive is active.
- If another start marker appears before the prospective end marker, Exempla uses the newer start marker.
- A passage can cross line breaks and speaker turns. Speaker identifies the person active where the returned passage begins.
- Include comments does not change a Between markers search. Exempla searches the literal source, and any speaker codes or parenthetical comments inside the returned passage are retained.
- Join at apostrophes is hidden because tokenization does not determine literal marker matching.
- Matches do not overlap. After Exempla closes one passage, it continues looking for the next start marker after that end marker.
- A long returned passage is shortened in the visible Token cell. Hover for an explanatory tooltip. The complete passage is retained for File View, the Exempla Clipboard, and CSV export.
Understand every search option before using it
Off by default. When off, Ottawa, OTTAWA, and ottawa are equivalent. Turn it on when uppercase and lowercase forms must match the query exactly.
Off by default. Turn it on to search text inside matching parentheses. A match found inside a comment displays (comment) in the Speaker column. Square-bracketed speaker codes remain excluded.
Off by default and shared across the project’s analytical tools. When off, we’re becomes the two tokens we and re. When on, it becomes one token.
For word and phrase searches, Exempla treats the ordinary apostrophe and common typographic apostrophes as equivalent. The visual character used in the source does not need to match the query. Between markers is literal, so its marker characters must match the source.
Read one row from the match outward
Begin with Token, then use the surrounding columns to establish interpretation, attribution, and location.
The code attributed to the occurrence. A comment match is labelled (comment). A result before any speaker code uses a dash.
The readable source text immediately before the match. It is right-aligned so the words nearest the Token are easiest to compare.
The complete matched word or phrase. For a long Between markers passage, the cell shows a shortened preview while retaining the complete passage internally.
The readable source text immediately after the match. Together with Left Context, it forms the keyword-in-context view.
The corpus text name containing the occurrence. It may differ from the original filename if you renamed the text in Exempla.
The Exempla transcript line where the match begins. If a phrase crosses a line boundary, this is the first line of the occurrence.
Marks the occurrence for toolbar actions such as Add to clipboard and Export selected results. It is distinct from highlighting a table row.
Adjust context, sort the table, and open the source
Change how much context appears
The button above the table initially reads Context · 80 characters/side. Open it to choose Words or Characters, enter the amount per side, and choose Apply.
Choose 20 to 2,000 characters per side. Exempla keeps the visible fragment readable and avoids ending it in the middle of a word.
Choose 1 to 200 countable tokens per side. The active apostrophe rule determines those token boundaries.
Returns the panel’s draft to 80 characters per side. It does not change the table until you choose Apply.
Runs the same query, scope, and search options again with the new context display. A very large search may therefore take time again.
Sort all results, not only the visible page
Choose the Speaker, Token, Source, or Line heading to sort the complete result set. The first click uses ascending order. The next click reverses it. Left Context, Right Context, and the checkbox heading are not sortable.
The status row reports the active choice, such as Speaker A–Z or Line High–Low. Sorting remains unavailable until the search is complete.
Choose Restore original order in the status row to return to the exact sequence produced by the search.
Open an occurrence in File View
Double-click a result row, press Return on a highlighted row, or right-click and choose Open in File View. Exempla opens the source transcript, moves to the recorded line, and highlights the matched word, phrase, or passage. This does not change Active selection.
Know exactly what each selection controls
Created by clicking, Shift-clicking, or Command-clicking on macOS and Ctrl-clicking on Windows. Highlighted rows control right-click commands, Copy, and Open in File View.
Created with the final-column checkboxes or the Select control. Checked results control the toolbar’s Add to clipboard and Selected results export actions, including selections on other pages.
Select one page or the complete result set
A result set with more than 1,000 occurrences is paginated. The bottom bar shows the page, the visible result range, and the complete count. Checked results remain checked when you move to another page.
- For a one-page result, Select all and Deselect all toggle the complete table directly.
- For multiple pages, Select opens choices for the current page, all results, deselecting the current page, and deselecting everything.
- The selected counter always refers to checked results across every page.
- Press Command+A on macOS or Ctrl+A on Windows to check every result in the completed result set.
- Press Space to check or uncheck the currently highlighted rows.
- Press Escape to remove row highlighting without changing checkboxes.
Send evidence to the Exempla Clipboard
Choose Add to clipboard to add the checked occurrences to the project’s persistent evidence table. Successfully added results are unchecked afterward. If the same result record is already present, Exempla skips the duplicate and reports how many results were added and skipped.
Select related occurrences from one or several Concordance pages.
Store them with their speaker, context, source, line, and complete Token value.
The collected evidence remains available after saving, closing, and reopening the project.
Use the right-click menu for direct row actions
Right-clicking the highlighted rows provides Open in File View, Add Selected to Exempla Clipboard, Add All Results to Exempla Clipboard, Copy Selected Results, Export Selected Results, Export All Results, and Select All. Actions requiring a completed search remain unavailable while the search is running.
Export checked results or the full Concordance
The Export menu above the table has two explicit choices.
Exports every checked occurrence across all pages. This option is unavailable until at least one result is checked.
Exports every occurrence in the completed result set, regardless of checkbox state. It is unavailable until the search finishes.
The file contains Speaker, Left Context, Token, Right Context, Source, and Line, followed by one row per exported occurrence. The checkbox column is not exported.
Exempla proposes a name containing the date, project name, Concordance description, and query when it fits safely on macOS and Windows.
The CSV uses UTF-8 with an encoding signature so accented characters open reliably in spreadsheet software on macOS and Windows.
Exports of at least 5,000 rows use a responsive progress window with Cancel. Exempla writes safely so cancelling does not leave a partial file in place.
Work safely with broad searches and millions of occurrences
A broad query such as Contains a can match millions of tokens. Exempla stores the result compactly and displays only the current page of up to 1,000 rows. Pagination limits the visual table; it does not reduce the actual result count.
The search runs outside the main interface. Use Cancel if the query is broader than intended. Cancelling discards the incomplete results.
Select all, sorting, full export, and Add All still refer to the complete result set, not only the 1,000 visible rows.
Hundreds of thousands of stored rows can use substantial working memory and make the project heavier. When an addition contains at least 100,000 results, or would bring the Clipboard to 100,000 rows, Exempla therefore recommends exporting the results directly. You can choose Export, Cancel, or Add Anyway.
A large addition uses progress and remains responsive. Cancelling is atomic, which means none of that attempted batch is added.
When a search does not return what you expected
Confirm that the project contains at least one text, the query is not empty, Active selection still contains valid items, and a Between markers query contains two markers.
Check your active selection, Match, Case-sensitive, Include comments, and Join at apostrophes. Then try a less restrictive scope or mode.
Confirm that its tokens are adjacent and in the same order. A phrase cannot cross into another speaker’s turn or skip an intervening word.
Turn on Include comments. Correctly parenthesized comments are excluded from ordinary Concordance searches by default.
Check whether Join at apostrophes is on. The query must agree with the project-wide token rule.
Contains with a short query can match many tokens. Cancel, choose Exact word or a longer query, or narrow Active selection.
Increase the Words or Characters value and choose Apply. Remember that applying context regenerates the same search.
Check results in the last column. Merely highlighting table rows does not activate the toolbar’s Selected results action.
Was this chapter helpful?
Your response will help improve the guide.