To define how English nouns and verbs are inflected.
- §1. Noun inflections
- §5. Plural inflections
- §14. Verb inflections
- §45. Present to past participles
- §46. Adjective agreements
§1. Noun inflections. The following trie looks at the start of a word, which we assume to be a noun, and decides whether to use the indefinite article "a" or "an". This is much more complicated than simply looking for a vowel as the first letter, as people often think until they try a few cases.
The following was compiled by Toby Nelson with the aid of a pronunciation dictionary and the "Official Scrabble Wordlist".
<singular-noun-to-its-indefinite-article> ::= <en-trie-indef-a> | <en-trie-indef-b> | <en-trie-indef-c>
- This is Preform grammar, not regular C code.
§2. Exceptions to the exceptions:
<en-trie-indef-a> ::= oneir* an | onero* an | ukiyo-e an | Japanese style of 17th-19th cent. printmaking urao* an | urial* an | uvarovite* an a rare emerald-green garnet, Ca3Cr2(SiO4)3
- This is Preform grammar, not regular C code.
<en-trie-indef-b> ::= eu* a | e.g., euphoria, eulogy ewe* a | female sheep ewftes a | Spens. form of an eft lizard ewghen a | made of yew, i.e., yewen ewk a | houri a | once* a | a Once and Future King one* a | but still use an for oneir- and onero- onst a | dialect form of once oui* a | e.g., a Ouija board or a ouistiti (a marmoset) u a | the letter U u-* a | e.g., U-boats u'* a | e.g., u's uakari a | the South American monkey ub* a | e.g., ubiquitous udal* a | udomet* a | uey a | colloquial for "U-turn", as in "he pulled a uey" ueys a | ufo* a | uganda* a | the country Uganda ugr* a | uint* a | uk* a | ulex a | the genus of gorse uli* a | ulo* a | ulu* a | una a | from "una corda", the musical term unabomb* a | the so-called Unabomber unalist a | unanimit* a | unanimous* a | unesco a | the United Nations cultural body unescos a | unia* a | unic* a | unif* a | unig* a | unil* a | unio* a | unip* a | uniq* a | unis* a | unit* a | univ* a | upas* a | ura* a | ure* a | uri* a | uru* a | usa* a | use* a | usi* a | usu* a | utas* a | ute* a | uti* a | uto* a | utr* a | uttoxeter* a | the English town of Uttoxeter uva* a | uvu* a
- This is Preform grammar, not regular C code.
§4. And finally the basic rules:
<en-trie-indef-c> ::= a* an | e* an | i* an | o* an | u* an | f an | f's an | f-* an | fbi an | fo an | frs an | h an | h's an | h-* an | e.g., H-bomb haute* an | e.g., haute cuisine, hauteur heir* an | hono* an | e.g., honorific, honorary doctorate hour* an | l an | l's an | l-* an | e.g., L-plate m an | m's an | m-* an | e.g., M-ration n an | n's an | n-* an | e.g., N-dimensional manifold r an | r's an | r-* an | rac an | Royal Automobile Club raf an | Royal Air Force rspca an | Royal Society for the Prevention of Cruelty to Animals rsvp an | s an | s's an | s-* an | x an | x's an | x-* an | e.g., X-ray xmas* an | yb* an | these are mostly obs., Spens., and/or arch. yc* an | yd* an | yf* an | yg* an | ym* an | yn* an | yp* an | e.g., ypsilon yr* an | ys* an | yt* an | e.g., Ytterbium, Yttrium yw* an
- This is Preform grammar, not regular C code.
§5. Plural inflections. The following takes a single word, assumes it to be a noun which meaningfully has a plural, and modifies it to the plural form. ("Golf" is a noun which doesn't sensibly have a plural; the algorithm here would return "golves".)
The trie here was derived from a partial implementation of Damian Conway's algorithm: see his paper "An Algorithmic Approach to English Pluralization", online at his website. The use of tries makes this somewhat faster than Conway's reference implementation, which for clarity's sake consists of a long sequence of regular-expression matches.
Conway divides plurals into modern and classical forms, and in cases where a noun has both, we take the modern form. Thus "phalanxes", not "phalanges". Because we focus on single words, we also omit prepositional phrases ("under water") and position names qualified by following adjectives ("procurator fiscal", "postmaster general"). Otherwise we omit only two cases, both involving capitalised proper nouns: nationality adjectives used as if they were nouns ("I saw two Japanese walking into the airport") and names of people used as if they were count nouns for a category of people like the one named ("We need more Wills, more Henrys.") — these are not likely to arise much in Inform usage, and they are awkward to implement with our tries because they depend on prefix as well as suffix and require case-dependency.
In its written form (as of November 2009, anyway), Conway's paper omits an important step from Algorithm 1, though it's present in his Perl implementation: the regular case of a sibilant suffix. (Ironically, this means that as stated Algorithm 1 pluralizes "suffix" incorrectly, as "suffixs".) I have filled this omission. I have also amended step 11, which considers the regular plural of a sibilant plus "o" suffix to include an "e", so that Conway produces "torsoes", "bozoes"; we will have "torsos" and "bozos".
<singular-noun-to-its-plural> ::= ... <en-trie-plural-uninflected> | ... <en-trie-plural-pronouns> | ... <en-trie-plural-irregular> | ... <en-trie-plural-irregular-inflections> | ... <en-trie-plural-assimilated-classical-inflections> | ... <en-trie-plural-irregular-o-suffixes> | ... <en-trie-plural-regular-inflections> | ... <en-trie-plural-append-s>
- This is Preform grammar, not regular C code.
§6. See Conway's table A.2. The following nouns, mostly names of kinds of animal, have the same plural as singular form: for example, chamois, salmon, goldfish.
<en-trie-plural-uninflected> ::= *fish 0 | *ois 0 | *sheep 0 | *deer 0 | *pox 0 | *itis 0 | bison 0 | flounder 0 | pliers 0 | bream 0 | gallows 0 | proceedings 0 | breeches 0 | graffiti 0 | rabies 0 | britches 0 | headquarters 0 | salmon 0 | carp 0 | herpes 0 | scissors 0 | chassis 0 | high-jinks 0 | sea-bass 0 | clippers 0 | homework 0 | series 0 | cod 0 | innings 0 | shears 0 | contretemps 0 | jackanapes 0 | species 0 | corps 0 | mackerel 0 | swine 0 | debris 0 | measles 0 | trout 0 | diabetes 0 | mews 0 | tuna 0 | djinn 0 | mumps 0 | whiting 0 | eland 0 | news 0 | wildebeest 0 | elk 0 | pincers 0
- This is Preform grammar, not regular C code.
§7. We may as well pluralise pronouns while we're at it.
<en-trie-plural-pronouns> ::= i we | you you | thou you | she they | he they | it they | they they | me us | you you | thee you | her them | him them | it them | them them | myself ourselves | yourself yourself | thyself yourself | herself themselves | himself themselves | itself themselves | themself themselves | oneself oneselves
- This is Preform grammar, not regular C code.
§8. We now reach Conway step 4. These are irregular plurals mostly coming from archaisms.
<en-trie-plural-irregular> ::= beef beefs | we neglect the classical "beeves" brother brothers | and "brethren" child children | cow cows | and "kine" ephemeris ephemerides | genie genies | and "genii" money moneys | and "monies" mongoose mongooses | mythos mythoi | octopus octopuses | and "octopodes" ox oxen | soliloquy soliloquies | trilby trilbys
- This is Preform grammar, not regular C code.
§9. Step 5. Now we reach a batch of irregular but fairly general inflected endings; for example, protozoon to protozoa, or metamorphosis to metamorphoses. Note that we differ from Conway in pluralizing blouse as blouses, not blice, and human as humans, not humen.
<en-trie-plural-irregular-inflections> ::= *human humans | Step 5 begins here *man 3men | *blouse 2ses | *louse 5lice | *mouse 5mice | *tooth 5teeth | *goose 5geese | *foot 4feet | *zoon 4zoa | *cis 3ces | *sis 3ses | *xis 3xes
- This is Preform grammar, not regular C code.
§10. Step 6. These are inflections from Latin and Greek which have survived into modern English:
<en-trie-plural-assimilated-classical-inflections> ::= alumna alumnae | from table A.10 alga algae | vertebra vertebrae | codex codices | from table A.14 murex murices | silex silices | aphelion aphelia | from table A.19 hyperbaton hyperbata | perihelion perihelia | asyndeton asyndeta | noumenon noumena | phenomenon phenomena | criterion criteria | organon organa | prolegomenon prolegomena | agendum agenda | from table A.20 datum data | extremum extrema | bacterium bacteria | desideratum desiderata | stratum strata | candelabrum candelabra | erratum errata | ovum ova
- This is Preform grammar, not regular C code.
§11. Step 11a. (We're not implementing Conway's steps in sequence: see below.) These -o endings are mostly loan words from Romance languages whose original inflections are assimilated.
<en-trie-plural-irregular-o-suffixes> ::= albino albinos | alto altos | archipelago archipelagos | armadillo armadillos | basso bassos | canto cantos | commando commandos | contralto contraltos | crescendo crescendos | ditto dittos | dynamo dynamos | embryo embryos | fiasco fiascos | generalissimo generalissimos | ghetto ghettos | guano guanos | inferno infernos | jumbo jumbos | lingo lingos | lumbago lumbagos | magneto magnetos | manifesto manifestos | medico medicos | octavo octavos | photo photos | pro pros | quarto quartos | rhino rhinos | solo solos | soprano sopranos | stylo stylos | tempo tempos
- This is Preform grammar, not regular C code.
§12. Conway steps 8 to 11. These are regular inflections depending only on word endings.
<en-trie-plural-regular-inflections> ::= *ch 0es | Step 8: "church" to "churches" *sh 0es | "rush" to "rushes" *ss 0es | "dress" to "dresses" *alf 1ves | Step 9: "calf" to "calves" *elf 1ves | "self" to "selves" *olf 1ves | "wolf" to "wolves" *eaf 1ves | "sheaf" to "sheaves" *arf 1ves | "wharf" to "wharves" *nife 2ves | "knife" to "knives" *life 2ves | "life" to "lives" *wife 2ves | "wife" to "wives" *ax 0es | Sibilant additions: "fax" to "faxes" *ex 0es | "sex" to "sexes" *ix 0es | "Weetabix" to "Weetabixes" *ox 0es | "fox" to "foxes" *ux 0es | "flux" to "fluxes" *as 0es | "gas" to "gases" *es 0es | *is 0es | "mantis" to "mantises" *os 0es | "thermos" to "thermoses" *us 0es | "abacus" to "abacuses" *az 0es | *ez 0es | "fez" to "fezes" *iz 0es | *oz 0es | *uz 0es | *ay 0s | Step 10 begins here *by 1ies | *cy 1ies | *dy 1ies | *ey 0s | *fy 1ies | *gy 1ies | *hy 1ies | *iy 0s | *jy 1ies | *ky 1ies | *ly 1ies | *my 1ies | *ny 1ies | *oy 0s | *py 1ies | *qy 1ies | *ry 1ies | *sy 1ies | *ty 1ies | *uy 0s | *vy 1ies | *wy 1ies | *xy 1ies | *yy 1ies | *zy 1ies | *ao 0s | Step 11b begins here *bo 1oes | *co 1oes | *do 1oes | *eo 0s | *fo 1oes | *go 1oes | *ho 1oes | *io 0s | *jo 1oes | *ko 1oes | *lo 1oes | *mo 1oes | *no 1oes | *oo 0s | *po 1oes | *qo 1oes | *ro 1oes | *so 0s | *to 1oes | *uo 0s | *vo 1oes | *wo 1oes | *xo 0s | *yo 1oes | *zo 0s
- This is Preform grammar, not regular C code.
§13. Lastly, the fallback if none of the above cases match: append an -s, of course.
<en-trie-plural-append-s> ::= * 0s Step 13
- This is Preform grammar, not regular C code.
§14. Verb inflections. "Le verbe est l'âme d'une langue" (attributed to Georges Duhamel). And the care of the soul is, of course, complicated. For example, the source text can say something like this:
The verb to flaunt means to wear.
This tells Inform that a new verb's infinitive is "flaunt", but not how to construct its other parts. We will use Preform grammar not only to define how to construct English verbs, but also in a way enabling it to be used with other languages too.
Inform uses five different tenses (present, past, present perfect, past perfect, and future), three persons, two numbers, two senses (true and false), and two voices (active and passive); in addition, it keeps track of the infinitive, past participle and present participle of a verb. Altogether that makes 123 potentially different versions of the original text. But of course there's a great deal of duplication in this, and almost all of the versions can be made using a much smaller number of genuinely different inflected versions of the word.
Our general strategy works like this:
- (a) Identify one or more verbs as being too irregular to fit into any pattern, and handle those as special cases.
- (b) For all other verbs, identify a set of inflected forms which covers all of the possibilities we need to make, and write a trie to handle each one.
- (c) Try to use a single conjugation to show how these forms are used, that is, how the different word forms map onto the possible tenses, persons, numbers, and so on.
§15. This gives us a certain amount of choice. What exactly is "too irregular"? In French, are all -er, -ir, and -re verbs "regular"? (Consider "aller", for example.) In English, it's possible to say that there are seven or so classes of verbs, all regular by their own standards; but most people say there's just one class of verb, and then irregular exceptions.
Our approach will follow Greenbaum, "Oxford English Grammar", at 4.14. Like Greenbaum, we will use the term "form type" for the different possible inflected versions of a verb word. The verb "to be" has eight form types (be, am, is, are, was, were, been, being), but it's unique in that respect — so this is one we will consider to be "too irregular", and will handle as a special case.
All other English verbs have five form types, though in many cases two or more of these have the same spelling. These we will number as follows, for reasons which will become clear below:
- (1) Infinitive: flaunt.
- (2) Present participle: flaunting.
- (3) Past participle: flaunted.
- (5) Third person singular present (or just "present"): flaunts.
- (6) Third person singular past (or just "past"): flaunted.
In regular verbs the past and past participle are the same, as they are here: he flaunted (past); he had flaunted (past participle). But English has around 600 commonly occurring irregular verbs in which they are different, sometimes unpredictably so: he went (past); he had gone (past participle). Irregularity sometimes makes these forms coincide rather than making them different: for example, to set has just three distinct forms — to set, he sets, he set, he had set, setting.
§16. Form types are numbered from 0 up to, potentially, a constant called MAX_FORM_TYPES. (This is so large that there shouldn't ever be need for more.) Form type 0 is always the original text, and is used as the basis from which the others are generated. For English verbs Inform always sets form type 0 to the infinitive, but this needn't be true if it's more natural in other languages to do something else.
We then reserve form types 1 to 3 for infinitive, present participle, and past participle, respectively, and this is required to be the case in all languages. Form type 4 is reserved for the "adjoint infinitive": if we are given the English base text "be able to see", for example, this will be recognised (see below) as "be able to" plus "see", and "see" will be the "adjoint infinitive". For most verbs, we won't use it.
That means that form types 5 and upward are free to be used as needed by each language. English needs two: the present (5) and past (6) forms.
define BASE_FORM_TYPE 0 define INFINITIVE_FORM_TYPE 1 define PRESENT_PARTICIPLE_FORM_TYPE 2 define PAST_PARTICIPLE_FORM_TYPE 3 define ADJOINT_INFINITIVE_FORM_TYPE 4 define MAX_FORM_TYPES 123
§17. We're now ready to write the <verb-conjugation-instructions>. This is a block which looks at the infinitive of the verb and decides which of several conjugations should be used. Badly irregular verbs get conjugations of their own, and others are grouped together. In French, for example, we might use this block of instructions to divide into different cases for -er, -ir, and -re verbs.
Each row takes the form of a pattern of words to match, followed by a nonterminal giving the conjugation to use if a match is made. Matches are literal except:
- (a) The tail ... means any string of one or more words, but can only be used as the tail. Any text matching it is written into the adjoint infinitive. So be able to ... matches "be able to touch" and sets the adjoint infinitive to "touch".
- (b) A pattern written in the form -xyz matches the tail of a verb. This isn't useful for English, but in French it neatly spots classes of verbs: for example, -er detects first-conjugation verbs such as "donner".
Note that we have to make sure every possible infinitive text matches at least one line, and the best way to ensure that is to finish up with ... as the last pattern — this matches anything.
§18. The instructions for English are quite concise, except for the presence of the awkward contracted informal forms of verbs. (These aren't used in Inform assertion sentences, but are needed for text substitutions.)
<verb-conjugation-instructions> ::= be <to-be-conjugation> | be able to ... <to-be-able-to-auxiliary> | be able to <to-be-able-to-conjugation> | could <modal-conjugation> | may <modal-conjugation> | might <modal-conjugation> | must <modal-conjugation> | should <modal-conjugation> | would <modal-conjugation> | auxiliary-have <to-have-conjugation> | do <to-do-conjugation> | 're <contracted-to-be-conjugation> | 've <contracted-to-have-conjugation> | aren't <arent-conjugation> | can't <cant-modal-conjugation> | don't <informal-negated-modal-conjugation> | haven't <informal-negated-modal-conjugation> | mayn't <informal-negated-modal-conjugation> | mightn't <informal-negated-modal-conjugation> | mustn't <informal-negated-modal-conjugation> | wouldn't <informal-negated-modal-conjugation> | couldn't <informal-negated-modal-conjugation> | shouldn't <informal-negated-modal-conjugation> | won't <informal-negated-modal-conjugation> | ... <regular-verb-conjugation>
- This is Preform grammar, not regular C code.
§19. We will start with two auxiliary verbs, that is, verbs used to construct forms of other verbs. The first is "to have"; as we'll see, English uses this to construct perfect tenses:
Peter has opened the gate. Jane had closed it.
"To have" doesn't really mean that anybody possessed anything here, except perhaps a history. It's simply used in conjunction with the past participle ("opened" and "closed") to form a tense. Verbs like this are called "auxiliary".
But it's not actually true, despite what concise grammars say, that English uses "to have" here; it uses a slight variation which differs in the negated forms. We write
I have not taken the lantern.
rather than
I do not have taken the lantern.
which strictly speaking ought to be correct. Inform handles this by using a modified form of "to have", which we'll call "to auxiliary-have", which differs only in its negative forms. We're only going to give this present and past tenses since it's never needed except as an auxiliary.
Anyway, this is an example of a "conjugation". The purpose of this is to set a few special verb forms — such as the present and past participles — and then give a recipe to make all of the many forms which the verb can take within sentences. The verb forms are numbered — see above — and the recipe is called a "tabulation". We'll specify the format for this below, when we get to a more complicated example, but briefly: this one sets the present participle (2) to "having", the past participle (3) to "had", and then names <to-have-tabulation> as the tabulation. The text doesn't have to be a single word, and some ingenious tricks are possible to form it from other verb forms; see below.
The marker <auxiliary-verb-only> means that this form of "have" can only be accessed from other verb conjugations, not via a text substitution for "[have]".
<to-have-conjugation> ::= 2 having | 3 had | <auxiliary-verb-only> | <not-instance-of-verb-at-run-time> | <to-have-tabulation>
- This is Preform grammar, not regular C code.
§20. Tabulations give instructions for how to construct 120 possible versions of the verb. These are divided up first into active and passive "voices":
Peter carries the lantern. [Active.]
The lantern is carried by Peter. [Passive.]
This makes two sets of 60. Each set contains five tenses, which in English are present (1), past (2), perfect (3), past perfect (4) and future (5).
Peter carries the lantern. [1]
Peter carried the lantern. [2]
Peter has carried the lantern. [3]
Peter had carried the lantern. [4]
Peter will carry the lantern. [5]
This makes five sets of 12. In each set there are six persons: first person singular, second person singular, third person singular, first person plural, second person plural, third person plural. We always write them in that order:
I carry the lantern. [1PS]
You carry the lantern. [2PS]
He carries the lantern. [3PS]
We carry the lantern. [1PP]
You [more than one person] carry the lantern. [2PP]
They carry the lantern. [3PP]
And that makes six sets of 2: the positive sense and the negative.
I carry the lantern. [Positive.]
I do not carry the lantern. [Negative]
To sum up, two voices times five tenses times six persons times two senses, which makes 120 versions in all.
A tabulation is best thought of as a short program to make these. Inform starts out with all 120 versions blank, and each tabulation step sets one or more versions. It's perfectly legal for later steps to override earlier ones; and it's legal to leave some versions unset, marking them not to be used. (We're going to ignore all of the passives and three of the active tenses, so we're only going to set 48 versions, in the case of auxiliary-to-have.)
Each step consists of a selector, followed by a text. The selector simply chooses which of the 120 forms to set. The selector always begins with "a" or "p", meaning active or passive; it can then optionally give a digit from 1 to 5, narrowing down to a given tense; and it can optionally give a plus or minus sign, narrowing down to positive or negative senses. In the following, for example, a2+ means active voice (a), past tense (2), positive (+). This nails down the selection to just 6 versions of the verb.
The text is used literally, except for the following:
- (a) The numbers 1, 2, 3, ..., expand into the verb forms with those numbers. For example, 2 expands into the present participle for the verb. If the number is followed by an open bracket, then an infinitive, then a close bracket, then it expands to the verb form for that verb. For example, the following expands to "sought":
3 ( seek )
- (b) Text in the form 1+xyz expands into verb form 1 but with the letters "xyz" added. For example, 1+ed for the verb "to mark" would expand to "marked", since 1 is the infinitive form. This feature is much more useful in heavily inflected languages like French.
- (c) If a bracket, an infinitive, then a close bracket, is given, it expands to the corresponding version of that verb. For example, the step a1+ ( grab ) back sets the positive present-tense versions of a verb to "I grab back", "you grab back", "he grabs back", and so on. Note that the matching persons are used, i.e., if we're expanding this to make the first person singular, we use the first person singular of the verb we're borrowing. Finally, we can change the tense by placing a tense marker inside the open brackets: a3+ ( t1 have ) grabbed sets the perfects to "I have grabbed", "you have grabbed", and so on — without the tense marker it would have been "I have have had grabbed", because "have" would expand to its perfect tense and not its present tense. The t1 means present tense; t2 means past tense, and so on.
- (d) If a nonterminal name is given, then it will be set of six texts; these are used for the six persons.
A simple example, then, which uses only feature (d) of these exotica:
<to-have-tabulation> ::= a1+ <to-have-present> | a1- <to-have-present> not | a2+ had | a2- had not
- This is Preform grammar, not regular C code.
§21. And this is an example of splitting into cases for the six persons, 1PS, 2PS, 3PS, 1PP, 2PP, 3PP. I have, you have, he has, we have, you have, they have. (This is more excitingly varied in other languages, of course.)
<to-have-present> ::= have | have | has | have | have | have
- This is Preform grammar, not regular C code.
§22. Next we have "to do", which is like "to have" in being fairly regular, as irregular verbs go. But we treat this as a special case because, again, we're going to need as an auxiliary verb when forming negatives ("Peter does not wear the hat" — note the "does not"). But this time we give the full treatment, creating all 60 active forms.
For the passive, though, we do something new. The selector p* is actually a way to set all 60 passive forms (which would normally be written p), but it tells Inform to use "to be" as an auxiliary. When we write the p* step:
p* done by
the effect is the same as writing:
p ( be ) done by
The difference is that Inform more efficiently implements the p* version, by implementing "done by" as if it were a preposition rather than as part of a verb. This parses more quickly and makes English passive forms play more nicely with implied uses of "to be". For example, in
number of things carried by the player (1)
Inform has to infer the meaning
number of things which are carried by the player (2)
and it can only do this if it recognises "carried by" as being prepositional in nature, like "on" or "in". In other words, if we wrote the p step above instead of the p* step, (2) would still work but (1) would not. (We may have to revisit this for languages other than English.)
<to-do-conjugation> ::= 2 doing | 3 done | <to-do-tabulation> <to-do-tabulation> ::= a1+ <to-do-present> | a1- <to-do-present> not | a2+ did | a2- did not | a3 ( t1 auxiliary-have ) done | a4 ( t2 auxiliary-have ) done | a5+ will do | a5- will not do | p* done by <to-do-present> ::= do | do | does | do | do | do
- This is Preform grammar, not regular C code.
§23. Regular English verbs, then, look like so. We will, for the first time, make heavy use of our numbered verb forms: for example, for the verb "to take", they would be "take" (1), "taking" (2), "taken" (3), "takes" (5) and "took" (6). We start with the infinitive ("take") in verb form 1, but (2), (3), (5) and (6) are initially blank — we have to make them somehow.
We do this by giving their definitions not as fixed wording, as we did for the verbs above, but as tries which act on the infinitive to produce a wording. For example, <en-trie-present-participle> is a trie which performs:
take --> taking
We will have to define these tries below. Note that the infinitive can consist of multiple words; if so, the first word is run through the tries, and the remaining words are left alone. For example, "grab onto" would be inflected to "grabs onto", "grabbing onto" and so on.
<regular-verb-conjugation> ::= 2 <en-trie-present-participle> | 3 <en-trie-past-participle> | 5 <en-trie-present-verb-form> | 6 <en-trie-past> | <regular-verb-tabulation>
- This is Preform grammar, not regular C code.
§24. Here we see our auxiliary verbs in use. For the negated present tense, "Peter does not carry the ball"; for the negated past tense, "Peter did not carry the ball" — in both cases, this is "to do" plus the infinitive "take". For the perfect tenses, "to have" plus the past participle — "Peter has carried the ball", "Peter had carried the ball". For the future tense, "will" plus the infinitive — "Peter will carry the ball". (We're actually not going to implement this as a verb because all its forms are just "will", and because "to will" also means "to leave a bequest".)
<regular-verb-tabulation> ::= a1+ <regular-verb-present> | a1- ( do ) 1 | a2+ 6 | a2- ( do ) 1 | a3 ( t1 auxiliary-have ) 3 | a4 ( t2 auxiliary-have ) 3 | a5+ will 1 | a5- will not 1 | p* 3 by
- This is Preform grammar, not regular C code.
§25. This looks odd, but what it says is that the present tense of a regular English verb is always the infinitive (I take, you take, we take, and so on) except for third person singular (he takes), which is different. (It's usually what the plural of the infinitive would be if the infinitive were a noun, as we'll see.)
<regular-verb-present> ::= 1 | 1 | 5 | 1 | 1 | 1
- This is Preform grammar, not regular C code.
§26. Now for our most irregular verb: "to be".
<to-be-conjugation> ::= 2 being | 3 been | <to-be-tabulation> <to-be-tabulation> ::= a1+ <to-be-present> | a1- <to-be-present> not | a2+ <to-be-past> | a2- <to-be-past> not | a3 ( t1 auxiliary-have ) been | a4 ( t2 auxiliary-have ) been | a5+ will be | a5- will not be <to-be-present> ::= am | are | is | are | are | are <to-be-past> ::= was | were | was | were | were | were
- This is Preform grammar, not regular C code.
§27. Except for tense formation (Peter "will" take the ball), the most common modal verb which can be used in Inform source text is "can". For example:
the number of people who can see the King
This is modal because it makes the seeing only a possibility, not an actuality. An awkward thing about modal verbs in English is that they are deficient, that is, not all their forms even exist. "Can" has no infinitive. ("To can" means to put food into a sealed metal container, which isn't the same thing at all.) "Can" also has no perfect or future tenses. On the other hand, it does have inflected present and past tenses, and we need to implement that. So we will invent the infinitive form "be able to", and make the verb from that, but using "can" and "could" instead of "is able to" and "was able to". "Can" is rather irregular as a verb: the third person singular doesn't inflect ("he can", not "he cans"), and the negative is written "cannot" instead of "can not", presumably because we find the two "n"s awkward to elide, so we always pronounce it that way and the spelling now follows.
<to-be-able-to-conjugation> ::= 2 <en-trie-present-participle> | 3 <en-trie-past-participle> | <to-be-able-to-tabulation> <to-be-able-to-tabulation> ::= a1+ can ++1 | a1- cannot ++1 | a2+ could ++1 | a2- could not ++1 | a3 ( t1 auxiliary-have ) been able to ++1 | a4 ( t2 auxiliary-have ) been able to ++1 | a5+ will be able to ++1 | a5- will not be able to ++1
- This is Preform grammar, not regular C code.
§28. Inform has only a simple understanding of what "can" means, so it doesn't allow the source text to use "can" in combination with arbitrary verbs. Instead, each legal combination has to be declared explicitly:
To be able to reach is a verb meaning ...
Inform implements all of this by passing "be able to reach" through the same verb-conjugation mechanisms as all other verbs ("take", "see", and so on). But at least the conjugation used is now simple. Recall that when the instructions grammar, right back at the start of this discussion of verbs, chooses which conjugation to use, it converts the text matching the wild-card ... into the "adjoint infinitive" form (4). We get to this conjugation by matching
be able to ...
so, for example, "be able to reach" results in 4 being set to "reach".
Note also the construction 3 ( 4 ) in the passive. The 3 means "take the past participle of the verb in brackets", and the 4 means that the text of this verb's infinitive is the contents of verb form 4. So, for example, for "be able to reach", 3 ( 4 ) expands to 3 ( reach ) which expands to "reached", and we get passive forms like "Peter can be reached by Jane".
<to-be-able-to-auxiliary> ::= 2 <en-trie-present-participle> | 3 <en-trie-past-participle> | <to-be-able-to-auxiliary-tabulation> <to-be-able-to-auxiliary-tabulation> ::= a ( be able to ) 4 | p ( be able to ) be 3 ( 4 ) by
- This is Preform grammar, not regular C code.
§29. The following handles the other English modal verbs ("might", "should" and so on) surprisingly easily. The notation ++1 means that the verb being modified should appear in verb form 1, and so on: for example, "might not lead" as "might not" plus form 1 of "to lead", i.e., "lead".
<modal-conjugation> ::= 2 <en-trie-present-participle> | 3 <en-trie-past-participle> | <modal-tabulation> <modal-tabulation> ::= a1+ 4 ++1 | a1- 4 not ++1 | a2+ 4 have ++2 | a2- 4 not have ++2 | a3+ 4 have ++2 | a3- 4 not have ++2 | a4+ 4 have ++2 | a4- 4 not have ++2 | a5+ 4 ++1 | a5- 4 not ++1
- This is Preform grammar, not regular C code.
§30. That completes our basic kit of verbs nicely. What's left is used only for generating text at run-time — for printing adaptive messages, that is; none of these oddball exceptional cases is otherwise used as a verb in Inform source text. None of them has any meaning to Inform.
Inform could fairly easily support the contractions "isn't", "aren't", "wasn't", "can't" and so on, but we've chosen not to do so. They save very little typing, and they greatly change the aesthetic style of Inform source text without changing its functionality. (If we allowed them, some authors would use them all the time, and other authors never, but others still would mix them incoherently.)
But we still want people to be able to write adaptive text which uses these contracted forms: otherwise, how could we write classic messages like
You can't go that way.
and have them adapt to other tenses and viewpoints?
First we'll tackle "to 's", the contracted form of "to be": I'm, you're, and so on. Exactly how these contractions are used in different tenses is something that varies with different dialects of English — for example, "you'll not take the ball" is now a little obsolete except in rural dialects — and we aren't even going to try to cope with that.
<contracted-to-be-conjugation> ::= 2 being | 3 been | <not-instance-of-verb-at-run-time> | <contracted-to-be-tabulation> <contracted-to-be-tabulation> ::= a1+ <contracted-to-be-present> | a1- <contracted-to-be-present> not | a2+ <contracted-to-be-past> | a2- <contracted-to-be-past-negated> | a3+ <contracted-to-have-present> been | a3- <contracted-to-have-present> not been | a4+ 'd been | a4- 'd not been | a5+ 'll be | a5- 'll not be <contracted-to-be-present> ::= 'm | 're | 's | 're | 're | 're <contracted-to-be-past> ::= was | were | was | were | were | were <contracted-to-be-past-negated> ::= wasn't | weren't | wasn't | weren't | weren't | weren't
- This is Preform grammar, not regular C code.
§31. And now "to 've", the contracted form of "to have". A subtle dialect point here concerns the negated present tense:
Sorry, I don't have a clue. [US]
Sorry, I haven't got a clue. [British]
Sorry, I haven't a clue. [British, but antiquated]
Sorry, I didn't have a clue. [US or British]
Sorry, I hadn't got a clue. [British]
But the American forms are becoming more common in British English, so we'll go with those.
<contracted-to-have-conjugation> ::= 2 having | 3 had | <not-instance-of-verb-at-run-time> | <contracted-to-have-tabulation> <contracted-to-have-tabulation> ::= a1+ <contracted-to-have-present> | a1- <contracted-to-have-present> not | a2+ had | a2- hadn't | a3+ <contracted-to-have-present> had | a3- <contracted-to-have-present> not had | a4+ 'd had | a4- 'd not had | a5+ 'll have | a5- 'll not have <contracted-to-have-present> ::= 've | 've | 's | 've | 've | 've
- This is Preform grammar, not regular C code.
§32. Now we come to "aren't", a negated form of "to be", but where the contraction occurs between the verb and the "not" rather than between the subject and the verb.
Again, Inform doesn't know or care what this means. We're simply going to teach it to conjugate it as if it were a verb in its own right. So "to aren't" will be conjugated "I am not", "you aren't", "he isn't", and so on. (We don't say "I amn't", possibly because the "mn" is too awkward, but possibly also because we'd more likely say "I'm not". Because this would make the spacing awkwardly difficult — we would need to backspace — we won't take that option here.)
<arent-conjugation> ::= 2 <en-trie-present-participle> | 3 <en-trie-past-participle> | <not-instance-of-verb-at-run-time> | <arent-tabulation> <arent-tabulation> ::= a1+ <arent-present> | a2+ <arent-past> | a3+ <arent-perfect> | a4+ hadn't been | a5+ won't be <arent-present> ::= am not | aren't | isn't | aren't | aren't | aren't <arent-past> ::= wasn't | weren't | wasn't | weren't | weren't | weren't <arent-perfect> ::= haven't been | haven't been | hasn't been | haven't been | haven't been | haven't been
- This is Preform grammar, not regular C code.
§33. And finally: the contracted informal negatives of various modal verbs which it's useful to be able to print, like the "can't" in
You can't go that way.
English has more modal verbs than one tends to remember, and the definition of "modal" itself arguable. This is the best we can do.
<informal-negated-modal-conjugation> ::= 2 <en-trie-present-participle> | 3 <en-trie-past-participle> | 5 <en-trie-modal-contracted-past> | 6 <en-trie-modal-contracted-future> | 7 <en-trie-modal-contracted-present> | <not-instance-of-verb-at-run-time> | <informal-negated-modal-tabulation> <informal-negated-modal-tabulation> ::= a1+ <informal-negated-modal-present> ++1 | a2+ 5 ++2 | a3+ 5 ++2 | a4+ 5 ++2 | a5+ 6 ++1 <informal-negated-modal-present> ::= 1 | 1 | 7 | 1 | 1 | 1
- This is Preform grammar, not regular C code.
§34. Together with special rules for can't, which is inevitably slightly different:
<cant-modal-conjugation> ::= 2 <en-trie-present-participle> | 3 <en-trie-past-participle> | <not-instance-of-verb-at-run-time> | <cant-modal-tabulation> <cant-modal-tabulation> ::= a1+ can't ++1 | a2+ couldn't ++1 | a3+ ( t1 haven't ) been able to ++1 | a4+ ( t2 haven't ) been able to ++1 | a5+ won't be able to ++1
- This is Preform grammar, not regular C code.
§35. We have special tries just to list the forms of the cases we will deal with. Tries can do fancy things (see below), but here they act just as a look-up table: for example, "won't" has present "won't", past "wouldn't" and future "won't".
Note that results of tries normally have to be single words; but that plus signs can be used if we absolutely have to introduce spaces.
<en-trie-modal-contracted-present> ::= can't can't | don't doesn't | haven't hasn't | won't won't | mayn't mayn't | mightn't mightn't | mustn't mustn't | wouldn't wouldn't | couldn't couldn't | shouldn't shouldn't <en-trie-modal-contracted-past> ::= can't couldn't | don't didn't | haven't hadn't | won't wouldn't | mayn't mayn't+have | mightn't mightn't+have | mustn't mustn't+have | wouldn't wouldn't+have | couldn't couldn't+have | shouldn't shouldn't+have <en-trie-modal-contracted-future> ::= can't won't+be+able+to | don't won't | haven't won't+have | won't won't | mayn't mayn't | mightn't mightn't | mustn't mustn't | wouldn't wouldn't | couldn't couldn't | shouldn't shouldn't
- This is Preform grammar, not regular C code.
§36. That's the end of the conjugations — the easy part, it turns out. We now need to create the four tries to make verb forms out of the infinitive: the present participle, the past participle, the third-person singular present tense, and the past tense.
We'll start with the present participle. This is actually quite hard, because in some cases it depends on pronunciation rather than spelling. Greenbaum's "Oxford English Grammar" summarises the general rules at 4.16, as follows:
- (a) If the base ends in -e but not -ee, -oe or -ye, drop the final -e before adding -ing: thus drive to driving, but see to seeing, dye to dyeing, and so on.
- (b) If the base ends in -ie, as well as dropping the -e, also change the -i to -y: thus die to dying, untie to untying.
- (c) If the base ends in a stressed syllable whose spelling ends with a single vowel and then a single consonant, then double the consonant before adding -ing. Thus tip to tipping (not tiping), but break to breaking (not breakking).
- (d) If the base ends in a vowel and then -c, add -king. This is not quite the same as consonant doubling and doesn't depend on the stress; thus mimic to mimicking, picnic to picnicking.
These are fairly clear-cut rules, though English doesn't enforce them in all cases, so that most dictionaries let you say either focusing or focussing, for example, and either gluing or glueing (note that rule (a) drops the -e from -ue endings, but it's not at all clear why this case should be different, which may be why people are doubtful here); and in America participles like traveling or programing or worshiping are allowed by some people (with -l, -m, -me, -p endings), but they aren't universal. Inform will stick to traditional English as described above.
The tricky thing is that (c) is really a phonetic rule, not a spelling rule. For example, we need to count a final -y and -w as vowels, not consonants, because that's what they sound like. But at least that can be read from the spelling, whereas the presence or absence of stress can't. An English word generally stresses just one syllable, and always stresses at least one, so a monosyllabic word is always stressed. With a polysyllabic word, there's no easy way to tell. Consider deter to deterring (stress on second syllable of deter), but meter to metering (stress on first syllable of meter).
§37. The following algorithm is due to Toby Nelson, who produced it from a dictionary of 14,689 English verbs, some of them quite obscure (to torpefy, anyone? to spuilzie? to cachinnate?). It's essentially a more detailed version of Greenbaum's rules above.
<en-trie-present-participle> ::= ... <en-trie-irregular-present-participle> | ... <en-trie-irregular-compound-present-participle> | ... <en-trie-regular-a-present-participle> | ... <en-trie-regular-b-present-participle> | ... <en-trie-regular-c-present-participle>
- This is Preform grammar, not regular C code.
§38. First of all there are some irregular cases — some for the usual suspects, but others for oddball verbs where English breaks the normal phonetic rules for the sake of clarity. For example, the participle of "singe" ought to be "singing", but in fact we write "singeing", purely to make it different from the act of producing a song.
<en-trie-irregular-present-participle> ::= boob 0ing | had 0ding | quad 0ding | quod 0ding | squid 0ding | whid 0ding | ballad 0ing | salad 0ing | invalid 0ing | ref 0fing | stravaig 0ing | scoog 0ing | scoug 0ing | yak 0king | yok 0king | lek 0king | trek 0king | spaniel 0ling | vermeil 0ling | madam 0ing | buckram 0ing | hem 0ming | emblem 0ing | item 0ing | slalom 0ing | alarum 0ing | possum 0ing | chalan 0ing | challan 0ing | tyran 0ning | den 0ning | hen 0ning | ken 0ning | misken 0ning | pen 0ning | unpen 0ning | sten 0ning | in 0ning | gin 0ning | begin 0ning | bin 0ning | sin 0ning | damaskin 0ing | trampolin 0ing | chagrin 0ing | satin 0ing | on 0ning | con 0ning | don 0ning | kon 0ning | fillip 0ing | turnip 0ing | sip 0ping | cop 0ping | lop 0ping | clop 0ping | flop 0ping | plop 0ping | slop 0ping | galop 0ping | up 0ping | cup 0ping | gar 0ring | mortar 0ing | sker 0ring | deter 0ring | inter 0ring | disinter 0ring | reinter 0ring | aver 0ring | abhor 0ring | vor 0ring | demur 0ring | fur 0ring | smur 0ring | caucus 0ing | sus 0sing | combat 0ing | ballat 0ing | curat 0ing | quadrat 0ing | bet 0ting | abet 0ting | fet 0ting | fidget 0ing | target 0ing | crochet 0ing | epithet 0ing | ratchet 0ing | let 0ting | blet 0ting | leaflet 0ting | relet 0ting | sublet 0ting | underlet 0ting | net 0ting | benet 0ting | overnet 0ting | pet 0ting | spet 0ting | ret 0ting | aret 0ting | fret 0ting | regret 0ting | basset 0ing | closet 0ing | corset 0ing | cosset 0ing | gusset 0ing | posset 0ing | roset 0ing | russet 0ing | briquet 0ting | coquet 0ting | duet 0ting | parquet 0ting | covet 0ing | unrivet 0ing | velvet 0ing | discomfit 0ing | profit 0ing | limit 0ing | delimit 0ing | vomit 0ing | rit 0ting | frit 0ting | grit 0ting | bit 0ting | dit 0ting | kit 0ting | sit 0ting | besit 0ting | outsit 0ting | resit 0ting | picot 0ing | ballot 0ing | pilot 0ing | parrot 0ing | debut 0ing | brut 0ing | div 0ing | ante 0ing | miscellaneous -e exceptions be 0ing | binge 0ing | birdie 0ing | centre 0ing | chasse 0ing | cicerone 0ing | dele 0ing | ensilage 0ing | facsimile 0ing | glace 0ing | jeelie 0ing | longe 0ing | lunge 0ing | ouglie 0ing | peenge 0ing | pie 0ing | quaere 0ing | queue 0ing | recce 0ing | route 0ing | reroute 0ing | restringe 0ing | saute 0eing | schappe 0ing | segue 0ing | singe 0ing | sortie 0ing | stymie 0ing | winge 0ing | swinge 0ing | tinge 0ing | unbe 0ing | vise 0ing | vogue 1ing | whinge 0ing | aleye 1ing | a few -ye exceptions baye 1ing | herye 1ing | nye 1ing | rallye 1ing | reaedifye 1ing | stye 1ing | undersaye 1ing
- This is Preform grammar, not regular C code.
§39. Now some exceptional forms where consonant doubling doesn't occur:
<en-trie-irregular-compound-present-participle> ::= *<gosyz>ie 0ing | e.g. boogieing *ae 0ing | e.g. spaeing *quit 0ting | acquitting, quitting, requitting *uret 0ting | carburetting, sulphuretting *budget 0ing | budgeting, underbudgeting *efer 0ring | deferring, preferring, referring *nfer 0ring | conferring, inferring *sfer 0ring | retransferring, transferring *bias 0sing | biassing, unbiassing *bishop 0ing | bishoping, unbishoping *woman 0ing | womaning, unwomaning *jambok 0king | jambokking, sjambokking *alog 0ing | dialoging, cataloging *daub 0ing daubing, bedaubing
- This is Preform grammar, not regular C code.
§40. And now rules for consonant doubling:
<en-trie-regular-a-present-participle> ::= *<aeiouy>b 0bing | *<dglmpw>ad 0ding | *<bhlnrtw>ed 0ding | *<bklr>id 0ding | *<cdghlnprst>od 0ding | *<bchmprtw>ud 0ding | *uf 0fing | *<aeiouy>g 0ging | *<bcdhiklmnprstuv>al 0ling | *<bcdfghkmnprstuvwz>el 0ling | *<cfmnrtv>il 0ling | *<bcrtv>ol 0ling | *<cn>ul 0ling | *<bcdghjlprw>am 0ming | *<glt>em 0ming | *<dhklnrw>im 0ming | *lom 0ming | *<bcghlmrstv>um 0ming | *<bcflmptvw>an 0ning | *<ry>en 0ning | *<dhklprtw>in 0ning | *<fw>on 0ning | *<dfghprst>un 0ning | *<cdghjlmnprstwyz>ap 0ping | *<klprt>ep 0ping | *<dhklnprtuyz>ip 0ping | *<bdhmprstuw>op 0ping | *<dhpst>up 0ping | *yp 0ping | *<bcfhjmnptw>ar 0ring | *<fhmst>ir 0ring | *dor 0ring | *<bclp>ur 0ring | *<bgmpv>as 0sing | *<mnrsu>es 0sing | *<hmpw>is 0sing | *<bcds>os 0sing | *<bclm>us 0sing | *<bcfhlmprtuvw>at 0ting | *<ghjstvw>et 0ting | *<fhlmnptw>it 0ting | *<bcdhjlnprstw>ot 0ting | *<bcghjlmnprt>ut 0ting | *<ei>v 0ving | *iz 0zing
- This is Preform grammar, not regular C code.
<en-trie-regular-b-present-participle> ::= *<aeiou>c 0king | *<eoy>e 0ing | *ie 2ying <en-trie-regular-c-present-participle> ::= *e 1ing | * 0ing
- This is Preform grammar, not regular C code.
§42. Next the past participle. As noted above, for most verbs this is the same as the past (e.g., he agreed and it was agreed); but there's a list of exceptions for Anglo-Saxon survivals (e.g., he chose and it was chosen). The exceptional cases were derived from Wikipedia's catalogue of irregular English verbs as it stood in May 2011, with a few archaisms and obscenities removed.
<en-trie-past-participle> ::= <en-trie-irregular-past-participle> | <en-trie-past> <en-trie-irregular-past-participle> ::= be been | have had | do did | arise arisen | awake awoken | bear borne | beat beaten | become become | befall befallen | beget begotten | begin begun | bespeak bespoken | bite bitten | blow blown | break broken | browbeat browbeaten | choose chosen | cleave cloven | come come | dive dived | draw drawn | drink drunk | drive driven | eat eaten | fall fallen | fly flown | forbear forborne | forbid forbidden | forego foregone | foreknow foreknown | forelie forlain | forerun forerun | foresee foreseen | forget forgotten | forgive forgiven | forgo forgone | forsake forsaken | forswear forsworn | freeze frozen | ghostwrite ghostwritten | give given | go gone | grow grown | hew hewn | hide hidden | interweave interwoven | know known | lade laden | misbecome misbecome | misbeget misbegotten | mischoose mischosen | misdo misdone | misget misgotten | misgive misgiven | misknow misknown | misshape misshapen | misspeak misspoken | mistake mistaken | miswrite miswritten | mow mown | outdo outdone | outgrow outgrown | outgrow outgrown | outrun outrun | outshine outshone | outswear outsworn | outthrow outthrown | overbear overborne | overblow overblown | overclothe overclad | overcome overcome | overdo overdone | overdraw overdrawn | overdrink overdrunk | overdrive overdriven | overeat overeaten | overfly overflown | overgrow overgrown | overlie overlain | override overridden | overrun overrun | oversee overseen | oversew oversewn | overshake overshaken | overstride overstridden | overtake overtaken | overwear overworn | overwrite overwritten | partake partaken | plead pled | redo redone | redraw redrawn | regrow regrown | rerun rerun | resing resung | retake retaken | retread retrodden | rewrite rewritten | ride ridden | ring rung | rise risen | rive riven | run run | saw sawn | see seen | sew sewn | shake shaken | shave shaven | shear shorn | shine shone | shoe shodden | show shown | shrink shrunk | shrive shriven | sing sung | sink sunk | slay slain | smite smitten | sow sown | speak spoken | spin spun | spit spit | spring sprung | steal stolen | stink stunk | stride stridden | bestride bestridden | strike stricken | strive striven | swear sworn | swell swollen | take taken | tear torn | thrive thriven | throw thrown | tread trodden | underbear underborne | underdo underdone | underdraw underdrawn | undergo undergone | undergrow undergrown | underrun underrun | undertake undertaken | underwrite underwritten | undo undone | wake woken | wear worn | weave woven | withdraw withdrawn | wring wrung | write written
- This is Preform grammar, not regular C code.
§43. That's the mandatory participles sorted out; so now we move on to the two additional verb forms used by English. First, the present form: a curiosity of English is that this is almost always formed as if it were the plural of the infinitive — thus "touch" becomes "touches". There are just a handful of exceptions to this.
<en-trie-present-verb-form> ::= <en-trie-irregular-third-person-present> | ... <singular-noun-to-its-plural> <en-trie-irregular-third-person-present> ::= be is | have has | do does
- This is Preform grammar, not regular C code.
§44. Second, the past. This is harder. Once again we have a catalogue of Anglo-Saxon past forms (e.g., he chose, not he chooses); and after those are out of the way, the rules are the same as for the present participle, except for adding -ed instead of -ing. The tricky part, again, is spotting when to double the consonant, which again depends on stress.
<en-trie-past> ::= ... <en-trie-irregular-past> | ... <en-trie-irregular-compound-past> | ... <en-trie-regular-a-past> | ... <en-trie-regular-b-past> | ... <en-trie-regular-c-past> <en-trie-irregular-past> ::= be was | do did | go went | in 0ned | on 0ned | up 0ped | bet bet | abet 0ted | bid bid | bin 0ned | bit 0ted | buy bought | con 0ned | cop 0ped | cup 0ped | cut cut | den 0ned | dig dug | dit 0ted | div 0ed | don 0ned | eat ate | fet 0ted | fit fitted | fly flew | fur 0red | gar 0red | get got | gin 0ned | had 0ded | hem 0med | hen 0ned | hit hit | ken 0ned | kit 0ted | kon 0ned | lay laid | lek 0ked | let let | let 0ted | lop 0ped | net 0ted | ante 0ed | nye 1ed | pay paid | pen penned | pet 0ted | pie 1ed | put put | ref 0fed | ret 0ted | aret 0ted | rid rid | rit 0ted | run ran | say said | see saw | set set | sin 0ned | sip 0ped | sit sat | sus 0sed | aver 0red | vor 0red | wed wedded | wet wetted | win won | yak 0ked | yok 0ked | baye 1ed | bear bore | beat beat | bend bent | abhor 0red | abide abided | bide bided | bind bound | bite bit | blet 0ted | blow blew | boob 0ed | brut 0ed | burn burnt | cast cast | clop 0ped | come came | deal dealt | dele 1ed | dive dove | drag dragged | draw drew | duet 0ted | fall fell | feed fed | feel felt | find found | flee fled | flop 0ped | fret 0ted | frit 0ted | give gave | grit 0ted | grow grew | hang hung | have had | hear heard | hide hid | hold held | hurt hurt | item 0ed | keep kept | knit knit | know knew | lade laded | lead led | lend lent | aleye 1ed | lose lost | make made | mean meant | meet met | plop 0ped | quad 0ded | quit quit | quod 0ded | read read | redo redid | rend rent | ride rode | ring rang | arise arose | rise rose | rive rove | seek sought | sell sold | send sent | shed shed | shoe shoed | shut shut | sing sang | sink sank | sker 0red | slip slipped | slit slit | slop 0ped | smur 0red | spet 0ted | spin span | spit spat | sten 0ned | stye 1ed | swim swam | take took | tear tore | tell told | trek 0ked | unbe 1ed | undo undid | vise 1ed | awake awoke | wake woke | wear wore | weep wept | whid 0ded | wind wound | beget begot | begin began | benet 0ted | beset beset | besit 0ted | binge 1ed | bleed bled | break broke | breed bred | bring brought | build built | burst burst | catch caught | cling clung | covet 0ed | creep crept | curat 0ed | debut 0ed | demur 0red | deter 0red | drink drank | drive drove | fight fought | fling flung | forgo forwent | galop 0ped | glace 0ed | grind ground | herye 1ed | hoise hoist | inlay inlaid | input input | inset inset | inter 0red | kneel knelt | alarum 0ed | leave left | light lit | limit 0ed | longe 1ed | lunge 1ed | madam 0ed | misdo misdid | outdo outdid | picot 0ed | pilot 0ed | prove proved | queue 1ed | reave reft | recce 1ed | recut recut | relet 0ted | repay repaid | rerun reran | reset reset | resit 0ted | roset 0ed | route 1ed | salad 0ed | satin 0ed | saute 0ed | scoog 0ed | scoug 0ed | segue 1ed | shake shook | shall should | shape shaped | shave shaved | shine shined | shoot shot | singe 1ed | sleep slept | slide slid | sling slung | slink slunk | smite smote | speak spoke | speed sped | spell spelt | spend spent | split split | squid 0ded | stand stood | steal stole | stick stuck | sting stung | stink stank | swear swore | sweep swept | swing swung | teach taught | think thought | throw threw | tinge 1ed | tread trod | tyran 0ned | unpen 0ned | unset unset | upset upset | vogue 1ed | vomit 0ed | weave wove | winge 1ed | worth worth | wring wrang | write wrote | ballad 0ed | ballat 0ed | ballot 0ed | basset 0ed | become became | befall befell | behold beheld | birdie 1ed | caucus 0ed | centre 1ed | chalan 0ed | chasse 1ed | choose chose | cleave clove | closet 0ed | clothe clothed | combat 0ed | coquet 0ted | corset 0ed | cosset 0ed | emblem 0ed | fidget 0ed | fillip 0ed | forbid forbade | forego forewent | forget forgot | freeze froze | gusset 0ed | jeelie 1ed | misfit misfitted | misget misgot | mishit mishit | misken 0ned | mislay mislaid | missay missaid | misset misset | mortar 0ed | naysay naysaid | ouglie 1ed | outbid outbid | output output | outrun outran | outsit outsat | overdo overdid | parrot 0ed | peenge 1ed | posset 0ed | possum 0ed | prepay prepaid | preset preset | profit 0ed | quaere 1ed | rallye 1ed | recast recast | redraw redrew | regret 0ted | regrow regrew | re-lay re-laid | remake remade | reread reread | resell resold | resend resent | resing resang | retake retook | retell retold | rewind rewound | russet 0ed | shrink shrank | shrive shrove | slalom 0ed | sortie 1ed | spread spread | spring sprang | stride strode | strike struck | string strung | strive strove | stymie 1ed | sublet sublet | sunset sunset | swinge 1ed | target 0ed | thrive throve | turnip 0ed | unbend unbent | unbind unbound | unhear unheard | unmake unmade | unwind unwound | uphold upheld | velvet 0ed | waylay waylaid | whinge 1ed | writhe writhed | beseech besought | bespeak bespoke | briquet 0ted | buckram 0ed | chagrin 0ed | challan 0ed | crochet 0ed | delimit 0ed | epithet 0ed | forbear forbore | forelay forelaid | forelie forlay | forerun foreran | foresee foresaw | forgive forgave | forsake forsook | gainsay gainsaid | inbreed inbred | invalid 0ed | leaflet 0ted | lipread lipread | miscast miscast | misdeal misdealt | misfeed misfed | misgive misgave | mishear misheard | mishold misheld | miskeep miskept | misknow misknew | mislead misled | misread misread | missend missent | mistake mistook | outgrow outgrew | outride outrode | outsell outsold | outswim outswam | outtell outtold | outwear outwore | overbid overbid | overbuy overbought | overeat overate | overfly overflew | overhit overhit | overlay overlaid | overlie overlay | overnet 0ted | overpay overpaid | overrun overran | oversee oversaw | overset overset | parquet 0ted | partake partook | podcast podcast | precast precast | quadrat 0ed | ratchet 0ed | rebuild rebuilt | reinter 0red | reroute 1ed | reshoot reshot | rethink rethought | retread retrod | rewrite rewrote | schappe 1ed | spaniel 0led | underdo underdid | undergo underwent | unrivet 0ed | vermeil 0led | webcast webcast | backbite backbit | bespread bespread | bestride bestrode | browbeat browbeat | cicerone 1ed | crosscut crosscut | damaskin 0ed | disinter 0red | ensilage 1ed | forecast forecast | foreknow foreknew | foretell foretold | forswear forswore | intercut intercut | misbeget misbegot | misshape misshaped | misshoot misshot | misspeak misspoke | misspell misspelt | misspend misspent | miswrite miswrote | outdrink outdrunk | outfight outfought | outshine outshone | outspend outspent | outswear outswore | outthink outthought | outthrow outthrew | overbear overbore | overbend overbent | overblow overblew | overcast overcast | overcome overcame | overdraw overdrew | overfeed overfed | overgrow overgrew | overhang overhung | overhear overheard | overlend overlent | override overrode | oversell oversold | overslip overslipped | overtake overtook | overwear overwore | sightsee sightsaw | stravaig 0ed | telecast telecast | unclothe unclothed | underbid underbid | underbuy underbought | undercut undercut | underdig underdug | underlay underlaid | underlet 0ted | underlie underlaid | underpay underpaid | underrun underran | unfreeze unfroze | withdraw withdrew | withhold withheld | broadcast broadcast | discomfit 0ed | facsimile 1ed | misbecome misbecame | mischoose mischose | outthrust outthrust | overbreed overbred | overbuild overbuilt | overdrink overdrank | overdrive overdrove | overshake overshook | overshine overshone | overshoot overshot | oversleep overslept | overslide overslid | overspend overspent | overswing overswung | overwrite overwrote | proofread proofread | reaedifye 1ed | restringe 1ed | simulcast simulcast | trampolin 0ed | underbear underbore | underbind underbound | undercast undercast | underdraw underdrew | underfeed underfed | undergrow undergrew | underhang underhung | undersaye 1ed | undersell undersold | undertake undertook | withstand withstood | ghostwrite ghostwrote | interbreed interbred | interweave interwove | overclothe overclothed | overstride overstrode | underbuild underbuilt | undershoot undershot | underspend underspent | understand understood | underwrite underwrote | underclothe underclothed | misunderstand misunderstood <en-trie-irregular-compound-past> ::= *<gosyz>ie 1ed | e.g. boogied *ae 1ed | e.g. spaed *quit 0ted | acquitted , quitted , requitted *uret 0ted | carburetted , sulphuretted *budget 0ed | budgeted , underbudgeted *efer 0red | deferred , preferred , referred *nfer 0red | conferred , inferred *sfer 0red | retransferred , transferred *bias 0sed | biassed , unbiassed *bishop 0ed | bishoped , unbishoped *woman 0ed | womaned , unwomaned *jambok 0ked | jambokked , sjambokked *alog 0ed | dialoged , cataloged *daub 0ed daubed , bedaubed <en-trie-regular-a-past> ::= *<aeiouy>b 0bed | *<dglmpw>ad 0ded | *<bhlnrtw>ed 0ded | *<bklr>id 0ded | *<cdghlnprst>od 0ded | *<bchmprtw>ud 0ded | *uf 0fed | *<aeiouy>g 0ged | *<bcdhiklmnprstuv>al 0led | *<bcdfghkmnprstuvwz>el 0led | *<cfmnrtv>il 0led | *<bcrtv>ol 0led | *<cn>ul 0led | *<bcdghjlprw>am 0med | *<glt>em 0med | *<dhklnrw>im 0med | *lom 0med | *<bcghlmrstv>um 0med | *<bcflmptvw>an 0ned | *<ry>en 0ned | *<dhklprtw>in 0ned | *<fw>on 0ned | *<dfghprst>un 0ned | *<cdghjlmnprstwyz>ap 0ped | *<klprt>ep 0ped | *<dhklnprtuyz>ip 0ped | *<bdhmprstuw>op 0ped | *<dhpst>up 0ped | *yp 0ped | *<bcfhjmnptw>ar 0red | *<fhmst>ir 0red | *dor 0red | *<bclp>ur 0red | *<bgmpv>as 0sed | *<mnrsu>es 0sed | *<hmpw>is 0sed | *<bcds>os 0sed | *<bclm>us 0sed | *<bcfhlmprtuvw>at 0ted | *<ghjstvw>et 0ted | *<fhlmnptw>it 0ted | *<bcdhjlnprstw>ot 0ted | *<bcghjlmnprt>ut 0ted | *<ei>v 0ved | *iz 0zed <en-trie-regular-b-past> ::= *<aeiou>c 0ked | magicked *<eioy>e 1ed | dried *<aeiou>y 0ed played, conveyed, convoyed, guyed, preyed <en-trie-regular-c-past> ::= *e 1ed | *y 1ied | shied, tried * 0ed
- This is Preform grammar, not regular C code.
§45. Present to past participles. Sentences like
The verb to carry means the carrying relation.
are only one way in which Inform creates new verbs; it also implicitly creates verbs when actions are declared:
Smoothing is an action applying to one thing.
Verbs like this are not stored in anything like the full conjugations above; the action knows just two forms of its verb, the present and past participles. This time the base text is the present participle ("smoothing"); the other inflection we need is the past participle ("smoothed"), and we need a trie which generates it from the present. This process is called "pasturising", which is, er, not actually an approved term from linguistics.
English is replete with exceptions — "catching" must become "caught", not "catched", for instance — so this trie consists of about 460 special cases followed by two general rules.
<pasturise-participle> ::= <en-trie-pasturise-exceptions> | ... <en-trie-pasturise-regular-y> | ... <en-trie-pasturise-regular> <en-trie-pasturise-exceptions> ::= abiding abided | alighting alighted | arising arisen | awaking awakened | backbiting backbitten | backsliding backslidden | bearing born | beating beaten | becoming become | befalling befallen | begeting begotten | begining begun | beholding beheld | bending bent | bereaving bereaved | beseeching besought | besetting beset | bestrewing bestrewn | betting bet | betaking betaken | bethinking bethought | binding bound | biting bitten | bleeding bled | blowing blown | breaking broken | breeding bred | bringing brought | broadcasting broadcast | browbeating browbeaten | building built | burning burned | bursting burst | busting busted | buying bought | casting cast | catching caught | chiding chided | choosing chosen | claping clapped | clinging clung | clothing clothed | coming come | costing cost | creeping crept | crossbreeding crossbred | cutting cut | daring dared | daydreaming daydreamed | dealing dealt | digging dug | dighting dighted | disproving disproved | diving dived | doing done | drawing drawn | dreaming dreamed | drinking drunk | driving driven | dwelling dwelt | eating eaten | enwinding enwound | falling fallen | feeding fed | feeling felt | fighting fought | finding found | fitting fitted | fleeing fled | flinging flung | flying flown | forbearing forborne | forbiding forbidden | fordoing fordone | forecasting forecast | foregoing foregone | foreknowing foreknown | foreruning forerun | foreseeing foreseen | foreshowing foreshown | forespeaking forespoken | foretelling foretold | forgetting forgotten | forgiving forgiven | forsaking forsaken | forswearing forsworn | fraughting fraught | freezing frozen | frostbiting frostbitten | gainsaying gainsaid | getting got | gilding gilded | giving given | going gone | grinding ground | growing grown | halterbreaking halterbroken | hamstringing hamstrung | hand-feeding hand-fed | handwriting handwritten | hanging hung | hearing heard | heaving heaved | hewing hewn | hiding hidden | hitting hit | holding held | hurting hurt | inbreeding inbred | inlaying inlaid | inputing input | insetting inset | interbreeding interbred | intercutting intercut | interlaying interlaid | intersetting interset | interweaving interwoven | interwinding interwound | inweaving inwoven | jerry-building jerry-built | keeping kept | kneeling knelt | knitting knitted | knowing known | lading laden | landsliding landslid | laying laid | leading led | leaning leaned | leaping leaped | learning learned | leaving left | lending lent | letting let | lieing lain | lighting lit | lip-reading lip-read | losing lost | making made | meaning meant | meeting met | misbecoming misbecome | miscasting miscast | miscutting miscut | misdealing misdealt | misdoing misdone | mishearing misheard | mishitting mishit | mislaying mislaid | misleading misled | mislearning mislearned | misreading misread | missaying missaid | missending missent | missetting misset | misspeaking misspoken | misspelling misspelled | misspending misspent | misswearing missworn | mistaking mistaken | misteaching mistaught | mistelling mistold | misthinking misthought | misunderstanding misunderstood | miswearing misworn | misweding miswed | miswriting miswritten | mowing mowed | offsetting offset | outbiding outbid | outbreeding outbred | outdoing outdone | outdrawing outdrawn | outdrinking outdrunk | outdriving outdriven | outfighting outfought | outflying outflown | outgrowing outgrown | outlaying outlaid | outleaping outleaped | outputing output | outriding outridden | outruning outrun | outseeing outseen | outselling outsold | outshining outshined | outshooting outshot | outsinging outsung | outsitting outsat | outsleeping outslept | outsmelling outsmelled | outspeaking outspoken | outspeeding outsped | outspending outspent | outspining outspun | outspringing outsprung | outstanding outstood | outswearing outsworn | outswiming outswum | outtelling outtold | outthinking outthought | outthrowing outthrown | outwearing outworn | outwinding outwound | outwriting outwritten | overbearing overborne | overbiding overbid | overbreeding overbred | overbuilding overbuilt | overbuying overbought | overcasting overcast | overcoming overcome | overcutting overcut | overdoing overdone | overdrawing overdrawn | overdrinking overdrunk | overeating overeaten | overfeeding overfed | overhanging overhung | overhearing overheard | overlaying overlaid | overleaping overleaped | overlieing overlain | overpaying overpaid | overriding overridden | overruning overrun | overseeing overseen | overselling oversold | oversetting overset | oversewing oversewn | overshooting overshot | oversleeping overslept | oversowing oversown | overspeaking overspoken | overspending overspent | overspilling overspilled | overspining overspun | overspreading overspread | overspringing oversprung | overstanding overstood | overstrewing overstrewn | overstriding overstridden | overstriking overstruck | overtaking overtaken | overthinking overthought | overthrowing overthrown | overwearing overworn | overwinding overwound | overwriting overwritten | partaking partaken | paying paid | pleading pleaded | praying prayed | prebuilding prebuilt | predoing predone | premaking premade | prepaying prepaid | preselling presold | presetting preset | preshrinking preshrunk | presplitting presplit | proofreading proofread | proving proven | putting put | quick-freezing quick-frozen | quiting quit | reading read | reawaking reawaken | rebiding rebid | rebinding rebound | rebroadcasting rebroadcast | rebuilding rebuilt | recasting recast | recutting recut | redealing redealt | redoing redone | redrawing redrawn | reeving reeved | refitting refitted | regrinding reground | regrowing regrown | rehanging rehung | rehearing reheard | reknitting reknitted | relearning relearned | relighting relit | remaking remade | rending rent | repaying repaid | rereading reread | reruning rerun | reselling resold | resending resent | resetting reset | resewing resewn | retaking retaken | reteaching retaught | retearing retorn | retelling retold | rethinking rethought | retreading retread | retrofitting retrofitted | rewaking rewaken | rewearing reworn | reweaving rewoven | reweding rewed | reweting rewet | rewining rewon | rewinding rewound | rewriting rewritten | riding rid | riding ridden | ringing rung | rising risen | riving riven | roughcasting roughcast | running run | sand-casting sand-cast | sawing sawed | saying said | seeing seen | seeking sought | self-feeding self-fed | self-sowing self-sown | selling sold | sending sent | setting set | sewing sewn | shaking shaken | shaving shaved | shearing sheared | sheding shed | shining shined | shoeing shoed | shooting shot | showing shown | shrinking shrunk | shriving shriven | shutting shut | sight-reading sight-read | singing sung | sinking sunk | siting sat | skywriting skywritten | sleeping slept | sliding slid | slinging slung | slinking slinked | slitting slit | smelling smelled | smiting smitten | sneaking sneaked | sowing sown | speaking spoken | speeding sped | spelling spelled | spending spent | spilling spilled | spinning spun | spitting spit | splitting split | spoiling spoiled | spoon-feeding spoon-fed | spreading spread | springing sprung | stall-feeding stall-fed | standing stood | staving staved | stealing stolen | sticking stuck | stinging stung | stinking stunk | strewing strewn | striding stridden | striking struck | stringing strung | striping stripped | striving striven | subletting sublet | sunburning sunburned | swearing sworn | sweating sweat | sweeping swept | swelling swollen | swiming swum | swinging swung | taking taken | teaching taught | tearing torn | telecasting telecast | telling told | test-driving test-driven | test-flying test-flown | thinking thought | thriving thrived | throwing thrown | thrusting thrust | treading trodden | troubleshooting troubleshot | typecasting typecast | typesetting typeset | typewriting typewritten | unbearing unborn | unbending unbent | unbinding unbound | unbuilding unbuilt | underbiding underbid | underbuying underbought | undercutting undercut | underfeeding underfed | undergoing undergone | underlaying underlaid | underletting underlet | underlieing underlain | underruning underrun | underselling undersold | undershooting undershot | underspending underspent | understanding understood | undertaking undertaken | underthrusting underthrust | underwriting underwritten | undoing undone | undrawing undrawn | unfreezing unfrozen | unhanging unhung | unhiding unhidden | unholding unheld | unknitting unknitted | unlading unladen | unlaying unlaid | unlearning unlearned | unmaking unmade | unreeving unreeved | unsaying unsaid | unsewing unsewn | unslinging unslung | unspining unspun | unsticking unstuck | unstringing unstrung | unswearing unsworn | unteaching untaught | unthinking unthought | unweaving unwoven | unwinding unwound | unwriting unwritten | upholding upheld | upsetting upset | vexing vexed | waking woken | waylaying waylaid | wearing worn | weaving woven | weding wed | weeping wept | wetting wet | wining won | winding wound | withdrawing withdrawn | withholding withheld | withstanding withstood | wringing wrung | writing written <en-trie-pasturise-regular-y> ::= *aying 3ed | e.g., "slaying" to "slayed" *eying 3ed | e.g., "preying" to "preyed" *oying 3ed | e.g., "toying" to "toyed" *ying 4ied e.g., "verifying" to "verified" <en-trie-pasturise-regular> ::= *ing 3ed e.g., "smashing" to "smashed"
- This is Preform grammar, not regular C code.
§46. Adjective agreements. English doesn't inflect adjectives at all (let's not argue about "blond" and "blonde"), so the following are just stubs.
<adjective-to-plural> ::= * 0 <adjective-to-masculine-singular> ::= * 0 <adjective-to-feminine-singular> ::= * 0 <adjective-to-masculine-plural> ::= * 0 <adjective-to-feminine-plural> ::= * 0
- This is Preform grammar, not regular C code.
§47. Grading of adjectives is more interesting. These spelling rules are taken from the Oxford English Grammar at 4.24, "Gradability and comparison". Something we can't easily implement is that a final vowel plus consonant doesn't result in doubling the consonant (in the way that "big" becomes "bigger") if that closing syllable is unstressed, but fortunately this is rare in English adjectives.
<adjective-to-comparative> ::= good better | well better | bad worse | far farther | *e 1er | e.g. "close" to "closer" *<bcdfghkmlnprstvwxyz>y 1ier | e.g. "ugly" to "uglier" *<aeiou><aeiou><bcdfghkmlnprstvxyz> 0er | e.g. "cheap" to "cheaper" not "cheapper" *<aeiou><bcdfghkmlnprstvxyz> 0+er | e.g. "fit" to "fitter" * 0er <adjective-to-superlative> ::= good best | well best | bad worst | far farthest | *e 1est | *<bcdfghkmlnprstvwxyz>y 1iest | *<aeiou><aeiou><bcdfghkmlnprstvxyz> 0est | *<aeiou><bcdfghkmlnprstvxyz> 0+est | * 0est
- This is Preform grammar, not regular C code.
§48. To the best of my knowledge there's no technical term for "the noun which is formed from an adjective to refer to the quality it measures", so the Inform source code calls this the "quiddity". English permits several competing forms of these to be constructed, depending on the adjective's spelling (for example, "brutal" can become "brutality", but "small" can't become "smallity"), but in general, except for Anglo-Saxon cases, the "-ness" suffix seems universally possible. For simplicity we'll use that; note the OEG's warning at 9.21 that this avoids problems where these forms, though notionally equivalent, have diverged in meaning: e.g., "casualty" should mean the same as "casualness", but no longer does. The "-ness" form is sometimes less elegant, but never means the wrong thing.
<adjective-to-quiddity> ::= *ong 3ength | e.g. "strong" to "strength" *<bcdfghkmlnprstvwxyz>y 1iness | e.g. "happy" to "happiness" * 0ness
- This is Preform grammar, not regular C code.
§49. English has almost no noun cases at all, with the only exceptions being Anglo-Saxon pronouns (thus we distinguish "they" and "them" as nominative and accusative, for example); and pronouns we handle separately in any case. We won't bother to distinguish gender:
<grammatical-case-names> ::= nominative | accusative <noun-declension> ::= * <en-noun-declension-group> <en-noun-declension-tables>
- This is Preform grammar, not regular C code.
§50. And the sorting into groups sorts everything into "group 1", the only group:
<en-noun-declension-group> ::= * 1 <en-noun-declension-tables> ::= <en-noun-declension-uninflected>
- This is Preform grammar, not regular C code.
§51. And in this single group, nominative and accusative forms are identical to the stem in both singular and plural.
<en-noun-declension-uninflected> ::= 0 | 0 | 0 | 0
- This is Preform grammar, not regular C code.
§52. English articles only inflect slightly, to show indefinite plurals; they don't distinguish nominative from accusative.
<article-declension> ::= * <en-article-declension> <en-article-declension> ::= a a a some some | the the the the the
- This is Preform grammar, not regular C code.