Template:pi-conj-special/documentation

Introduction
This template presents the conjugation of a Pali verb. It can include all 8 tenses and moods conjugated for person. (The word 'special' in the name is now a misnomer.) It also includes the present and future participles, which are formed by generally regular rules from the tenses. The forms of the verb are specified in two ways - by stems and by entry of the individual forms. It is also possible to suppress the derivation of the forms for specified combinations of person, tense and voice.

The processing to generate the tables is primarily in.

Labelling
The table is labelled by the provided item that is earliest in the list ...
 * label
 * positional parameter 1
 * cite
 * cite2
 * cite3
 * The name of the page

Exclusions
The passive, causatives, desideratives and intensives are treated as separate, lexically derived verbs.

The past participle, absolutive and gerundive are not regularly derived from the finite stems, and are therefore not handled by this template.

Grouping of Stems
The tenses and moods come in four groups, which are assigned their own stems:


 * 1) The 'special' stem(s), from which are formed the present tense, the (rare) imperfect tense, the imperative, and the optative. These stems are input in the 3s of the present.
 * 2) The aorist stem(s), from which the aorist is formed. These stems are input in the 3s of the aorist active.
 * 3) The perfect stem(s), from which the perfect tense is formed. At present, there is no way to input a perfect stem.
 * 4) The future stem, from which the future tense and conditional mood are formed. These stems are input in the 3s of the future.

Supplementary stems can be added for the imperfect tense, the optative mood and the conditional mood. Supplementary stems are needed for the imperfect and the conditional because of the vagaries of the mode of inclusion of the augment; its nominal position is between the prefix and the stem, but the prefix and the stem may have assimilated, and the first consonant of the stem may have degeminated. The optative has maintained a degree of independence from the present tense, and therefore freedom is need to modify the set of stems used. Additionally, a pattern of strong stem in the active and weak stem in the middle breaks down in the optative.

Division of Stems by Voice
The special and future stems can each be organised in one of four manners:


 * 1) A set common to both voices. These stems are categorised as both the 'active stems' and as the 'middle stems'.
 * 2) Separate sets for the active and middle voices. This accommodates the verbs that use a 'strong' form of the tense in the active and a 'weak' form in the middle, as with  for the active and kurute in the middle.  The two sets are respectively categorised as the 'active stems' and the 'middle stems'.
 * 3) Only stems for the active voice. For many verbs, the middle voice is only attested scantily or is completely absent.  These stems are categorised as the 'active stems'.
 * 4) Only stems for the middle voice. This is of limited use, though some grammars have falsely asserted that passive verbs are only used in the middle voice.  These stems are categorised as the 'middle stems'

The only set, or else the set used for the active voice, is referred to below as the first set. If there is another set, used only for the middle voice, it is referred to below as the second set.

The various forms are formed from the 'special' stems as follows:

The various forms are formed from the 'future' stems as follows:

Input of Present Stems
The stems are input in the 3s of the present tense. A stem supplied as a middle voice form will only be used for generating forms in the middle voice; a stem supplied in an active voice form will be used for both voices if not suppressed by midl or act.

The first set of stems is supplied as the sole and optional positional parameter and by the keyword cite. If neither is present, the first set of stems will consist of the lemma itself. The second set, restricted to use in generating middle forms, is introduced by the keyword midl.

Pali verbs often have alternative forms. Alternative forms of the stems may be specified using keywords cite2, cite3 and so on and parameters midl2, midl3 and so on. The forms specified for the first should all be active or all be middle.

Tenses and Voices to Display
The generation of middle forms from special stems is suppressed by adding the argument act. If the first member of the first set is middle, the generation of active forms will be suppressed.

For example, to display forms of coreti with only the contracted forms in the active, one could specify coreti. To display only the active, contracted forms, one could specify coreti. To display all the active forms, but only active forms, one could specify coreti.

Likewise, the generation of middle forms from future stems is suppressed by adding the argument act. If the first member of the first set is middle, the generation of active forms will be suppressed.

Which voices have forms generated from stems is explicitly defined by the parameters voice for the present, impf_voice for the imperfect, impr_voice for the imperative, aori_voice for the optative, futu_voice for the future, cond_voice for the conditional, part_voice for the present participles and futu_part_voice for the future participles. The valid values are 'both', 'act' to display only the active, 'midl' to display only the middle, and 'none' for no display. A value is ascribed to the parameter voice from the form of the active or only stem when not given by act or none. The explicit setting midl is disallowed. If the form is active, voice is ascribed the value 'both'; otherwise it is ascribed the value 'midl'.

Note that if no forms are generated from stems for a tense or mood, and no individual forms are supplied for it, there will be no displays for it. Unpopulated persons (tense by tense) and voices are likewise not displayed.

Individual Forms
Individual forms of the verb may be supplemented or replaced by a parameter expressing voice (a/m), person (1/2/3) and number (s/p), e.g., which yields the present active

Multiple supplementary forms could be specified by parameters a3p2, a3p3 and so on. The present active and middle participles can similarly be supplemented using 'pap' and 'pmp' respectively, and the future participles by fap and fmp.

The precise effect is specified by another parameter, named by suffixing '_mod', which has the following effects:

If there are multiple supplemental forms, this parameter applies to all of them; there is no parameter a3p2_mod.

For the other tenses and moods, one prefixes 'impf_' for the imperfect, 'impr_' for the imperative, 'opta_' for the optative, 'aori_' for the aorist, 'perf_' for the perfect, 'futu_' for the future and 'cond_' for the conditional.

Tuning of Writing System
For the Thai and Lao scripts, a fundamental difference is whether the writing system uses implicit vowels or not. The use of implicit vowels can by selected by no. By default, they are used; this selection can be confirmed by yes.

Some Lao script writing systems use ຍ for Pali  while others use ຢ. The option can be selected using the parameter y. ຍ can be selected by ຍ or yung. ຢ can be selected by ຢ or yaa. Both will be used if the option both is specified. The default is the default transliteration from, currently ຍ.

In the Burmese and Tai Tham scripts, there are two different vowels corresponding to ā. The choice depends on the preceding consonants, but there are different conventions. The parameter aa selects which value is used in the inflectional ending if the attested conventions may make different choices. The valid values are 'round', 'tall', 'both' and 'default'.

Note that the stem may limit the number of applicable conventions; it is the responsibility of the editor to avoid impossible combinations.

So far, aa has only been implemented for the Tai Tham script. The implementation is actually in Module:pi-decl/noun, so that the logic may be shared with nominal inflection.

Imperfect Tense
The imperfect has, in principle, an optional augment. The behaviour of the augment is defined by the parameter 'augment' as follows:

If the call has no midl parameter, the augmented form is supplied by the parameter impf. Otherwise, the augmented form is supplied by the parameter impf for the active and by impf_midl for the middle. If there are multiple forms, the extra forms are supplied in parameters impf2, impf3 and so on and impf_midl2 and impf_midl3 and so on. It is considered an error for impf_midl to be supplied although midl is not supplied.

The value of impf_voice defaults to none.

Imperative Mood
The value of impr_voice defaults to the assigned value of voice.

Optative Mood
For the optative active, the first group of stems is replaced or supplemented by the parameter opta; the manner is specified by parameter opta_mod in the same way as for fine-tuning. The option blank is treated as equivalent to none. If the parameter opta_midl is supplied, the forms it supplies (as third-singular optative middle forms) are the only ones used for the optative middle; otherwise the same stems are used as for the optative active. Additional replacement or supplementary forms can be specified using opta2 and opta_midl2 and so on.

The third-singular forms supplied by parameters opta and opta_midl must end in -eyya, -ā (e.g. for ), or -etha. The short singular active optatives in -e will not be generated from these optative forms; if appropriate, they must be added as fine-tuning.

The value of opta_voice derives from that assigned to voice.

Aorist Tense
The parameters aori_augment and aori_voice been created for future expansion, but they should not be used.

Aorist stems are input by aor2 and aori, with additional stems supplied by aor22 and aori2 and so on. The stems input via aor2 etc. are duplicated, giving one without the augment and another with the augment prefixed. Stems input by aori are not modified for the augment. Note that only active tense forms are generated from these stems.

Four aorist suffixes are recognised: -esi, -āsi (but only assuredly if 1), -ayi and -i, with paradiɡms as followsː

Note that the 1p and 2p of adesesi are missing. There presence seems to vary from verb to verb - and might be supplied by the corresponding forms of adesayi! There may need to be revisions to these paradigms. The 3rd plural form of pivi may not be assured for verbs otherwise conjugated the same and the 1st plural in -mhā might not be attested well enough. If you have attestations of these forms, it would be as well to record them as individual forms.

It is necessary to have the flag aasi to constrain the selection of the special -āsi paradigm. Many verbs with aorists ending in -āsi are conjugated like pivi, e.g. . Verbs following the special, contracting paradigm are recorded in category:Pali verbs with aorist in -āsi/-a.

Perfect Tense
The value of perf_voice defaults to 'both', but as perfect stems are currently not supported, the default is irrelevant.

Future Tense
The future tense, conditional mood and future participles will not be displayed unless the parameter futu is supplied.

Input of Stems
The stems are input in the 3s of the future tense. A stem supplied as a middle voice form will only be used for generating forms in the middle voice; a stem supplied in an active voice form will be used for both voices if not suppressed by futu_midl or act.

The first set of stems is supplied via the keyword futu. The second set, restricted to use in generating middle forms, is introduced by the keyword futu_midl.

Alternative forms of the stems may be specified using keywords futu2, futu3 and so on and parameters futu_midl2, futu_midl3 and so on. The forms specified for the first set should all be active or all be middle.

The choice of stems for each of the future and conditional is explained above.

The valid input values for futu_voice are 'both', 'act' and 'none'. It defaults to 'both', and the value 'midl' will be derived if the value supplied for futu is a middle form rather than an active form.

Examples
For example, to display active forms for and  with only the middle forms of, one can use, along with the parameters used to remove the system of the present, :

It would be more natural to use karissate, but for convenience the parameter will accept either an active or a middle form.

To display only the active forms, one can use :

Conditional Mood
As with the imperfect, the conditional has, in principle, an optional augment. The handling is is defined by the parameter cond_augment as follows:

If the verb has no futu_midl parameter, the augmented form is supplied by the parameter cond. Otherwise, the augmented form is supplied by the parameter cond for the active and by the parameter cond_midl for the middle. If there are multiple forms, the extra forms are supplied in parameters cond2, cond3 and so on and cond_midl2 and cond_midl3 and so on. It is considered an error for cond_midl to be supplied although futu_midl is not supplied.

There is a theoretical possibility that with implicit vowels a conditional 3s could be ambiguous between an active form ending in -ttha and a middle form ending in -atha. This can be avoided by instead using the active form in -tthā, which will have the same effect on the table.

The defaulting of cond_augment is slightly tricky. If argument has a value beginning with 'with', that is the default. Otherwise, it defaults to 'with_made'. The logic is that a value starting 'with' reflects the morphology and phonology of the stem, and therefore such a setting is likely to be appropriate for the conditional.

Parameter cond_voice defaults to the value assigned to futu_voice.

Problems with the Conditional
Grammars give discordant accounts of the terminations of the conditional. To resolve these issues, it seems worthwhile recording each attested conditional form in a citation for future analysis. Verbs with conditionals in their inflection tables are being recorded in the maintenance category category:Pali verbs exhibiting conditional mood.

There is a particular problem with the ending of the first person plural active - some authorities say -amhā, but others say -āma. The template leaves that entry blank; the blank should only be filled with an attested form. Verbs with such manual tuning are being recorded in the maintenance category category:Pali verbs with manual 1p conditional. There is a very real possibility that this may now have become a regional difference, like the choice between initial vy and by.

Example
To display what is claimed as the optative of as a conditional, one could use the template call .

.

Non-Roman Scripts
The standard nine non-Roman scripts are supported. The conjugation table is produced in a single script, one of the ten supported. The augment and suffixes are generated in that script, and their juncture takes writing conventions into account. (One omission is that having different ways of writing 'a' in closed and open syllables is not taken into account.) As noted above, various options are available to narrow or generalise the writing system of generated forms as appropriate.

Non-standard abugidas are also supported for inflection, provided there are no non-standard interactions between stem and ending. The following need to be provided to generate tables for such writing systems:
 * Transliteration to the abugida in Module:pi-Latn-translit
 * Transliteration from the abugida by the transliteration function listed in Module:languages/data/2.
 * Use of sc in the template invocation to specify the abugida.

Transliteration
By default, the transliteration of verb forms are also shown. There are two options to fine tune the behaviour of this transliteration: