Updating Unicorn's Greek Dictionary
This page documents the procedure for updating Unicorn's Greek Dictionary. It does not describe how to use the Greek dictionary.
In most cases imitating existing entries will work just fine. The “Test Dictionary” and “Test Selection” tools described below will catch many simple problems.
If you do add entries to your copy of the dictionary file, please send those entries back for inclusion in the dictionary distributed with Unicorn.
The Format of Entries
An entry is a single line divided into two or three sections by double colons '::'. The first section contains the various comma-separated forms of a word. The second section is the word's definition and any comments such as usage. Some entries have a third section, the metadata section, containing additional information used by the dictionary program to analyze the word's forms. A line starting with a sharp sign '#' is a comment line and will be ignored.
The layout of the forms section of an entry is simple. A noun entry explicitly lists the nominative and genitive singular forms and the nominative singular article. A verb entry calls out all six principal parts. An adjective entry has either two or three principal parts. For all parts of speech except the noun, use a vertical bar '|' to separate any variants of a form. A missing form is indicated with a single hyphen '-'.
A doubled semicolon ';;' is used in a definition section as a formatting directive. It tells the dictionary display to break a line and start a new, indented line. Only one of the two semicolons will be printed.
Metadata directives may be a single word (for example, prefix) or a pair of words separated by a colon (for example, imperfect-stem:εἶχ). If an entry has multiple metadata directives, they are separated by commas.
Parts of Speech
Greek is a less than regular language. The following paragraphs describe some subtleties in setting up dictionary entries, so that all forms will be correctly analyzed.
For some adjectives the genitive singular is needed and so is appended to the nominative singular with an at-sign '@'. For first and second declension contracted adjectives, you should list the contracted form with its definition. The uncontracted form may be listed in a separate entry, its definition referring to the main entry. See the examples of χρυσοῦς and χρυσέος. Adjectives whose neuter nominative form may end in either -ο or -ον should be listed under their -ο form. Metadata directives for adjectives include comp for the masculine nominative singular form of the (irregular) comparative and super for the masculine nominative singular form of the (irregular) superlative. See μεγας for an example.
A compound verb is indicated by putting a hyphen between the prefix and the main verb in the first principal part. There must also be prefix entries for any prefixes used with compound verbs. Accents should be properly placed on a verb's principal parts. The following are metadata directives for verbs:
- unaugmented-aorist-stem:stem notes an irregular unaugmented aorist stem; see αἰρεω.
- imperfect-stem:stem notes an irregular imperfect stem; see ἔχω.
- eta-contract flags verbs such as ζαω and χραομαι that contract to eta rather than long alpha.
- future-alpha-contract flags verbs such as ἐλαυνω whose future contracts with alpha.
- double-augment indicates a compound verb takes a double augment, that is, both the stem and the prefix take an augment,.
- adjective:adjective specified the masculine nominative singular of an irregularly formed verbal adjective.
- pluperfect-middle-is-unaugmented indicates that the pluperfect middle/passive does not take an augment; see ἐλέγχω.
- second-perfect indicates that the verb has additional second perfect forms. It drops the two letters of suffix from the perfect active principal part. See βαινω and its compounds.
- perfect-active-participle-stem:stem is used in conjunction with second-perfect to specify the irregularly formed stem of the perfect active participle.
- no-augment flags a verb without any augments, for example, ἀεροδρομέω.
A prefix is indicated by putting the keyword prefix in the metadata portion of the entry, for example
ὑπερ-::over; greatly; on behalf of::prefix ἀπο-|ἀπ-::away from ::prefix ἐπι-|ἐπ-|ἐφ-::upon, over, against, after::prefix
Any second alternate form of a prefix is used in front of a stem that begins with a vowel. A third alternative form is used in from of a stem that begins with an aspiration.
The prefix entries in the dictionary file should come before any verb entries.
A suffix is indicated by putting the keyword suffix in the metadata portion of the entry. Examples of suffixes include -περ, γε, and ουν.
It is not possible to specify alternate forms for nouns using the '|' character. Instead use one of the following metadata directives:
- vocative-sing, vocative-pl
- nominative-sing, nominative-pl
- genitive-sing, genitive-pl
- dative-sing, dative-pl
- accusative-sing, accusative-pl
- nav-dual, gd-dual
Words formed from the crasis of two other words should include a note such as (crasis of X and Y) in the definition. The keyword crasis is used by the dictionary to prevent the sanity checker from complaining about a breathing mark in the middle of the word.
Some words are so irregular that their entries must be generated by the dictionary program itself. When such entries appear in the dictionary, they are preceded by the comment character '#' and are marked with the metadata directive internal. Such entries can not be edited.
Tools for Adding Greek Dictionary Entries
In the Advanced pane of the Preferences dialog is a checkbox labeled “Enable special menu for Greek dictionary work”. When checked, a menu labeled “Special” will appear. Its menu items are useful tools for managing the Greek dictionary. The Latin dictionary does not have such tools at this point.
The “Dump Dictionary” menu item dumps a sorted copy of the dictionary into a new window.
The “Test Dictionary” menu item runs a set of sanity checks over the dictionary and reports the results in a new window. It performs the following actions:
- checks that a compound verb has the same sort of metadata as the base verb,
- reports any duplicates, unless they are flagged with a not-a-duplicate metadata directive,
- checks that each part of a word has an accent mark, unless the word enclitic or proclitic appears in the word's definition,
- checks for the right number and placement of breathing marks (note that including the word crasis in the definition of a crasis form will suppress complaints),
- checks for spurious non-Greek characters in the word,
- and checks that each principal part of the word can be looked up in the dictionary.
The “Test Selection” menu item runs the sanity checks described above over a region of selected text and reports the results in a new window. This command is typically used when creating a set of new entries for eventual inclusion in the dictionary. You can not use “Test Selection” on the dictionary file itself.
“Report Changes” is used for merging dictionary files. The dictionary loaded by Unicorn is considered the base dictionary. Any highlighted text is considered a set of changes to that base dictionary. Highlighting a window containing a file of changes to the base dictionary would be a typical scenario. Selecting “Report Changes” will first run the sanity checks and report any errors in a new window. If there are no errors, two new windows will opened, one containing altered entries (the word exists in the base dictionary, but has a different definition or metadata section) and the other containing new entries, not present in the base dictionary.
The final two menu items are used for determining the six principal parts of two common types of verbs. For both you highlight one or more lines of text, then select the desired menu item. If successful, a line of text will be replaced by the six principal parts of a verb, otherwise the line will be left unaltered. The “Compound Verb” menu item will take a compound verb with any set of endings. The “Pure Verb” menu item takes the first principal part of a thematic verb whose stem ends in a vowel.