Verb Phrases

19.1 The Structure of the Verb Phrase

The Alashian verb phrase has the following overall structure:

[pronominal clitic] [auxiliary verb] [main verb] [secondary verb] [pronominal clitic]

No verb, however, employs all slots at once; the vast majority only use two or three at once. The only slot that is filled for every verb is of course the main verb.

The secondary verb slot is used in multi-verb phrases, as in “want to X” or “forbid to X”. It is typically filled by an infinitive, usually preceded by some sort of marker such as the preposition λι- li-.

Most Alashian verbs require a clitic pronoun of some sort; whether it goes before or after the main verb depends on the morphological form of the main verb. Clitic pronouns may be absent only in a few specific cases, such as objectless imperatives, reflexive imperatives, and occasionally when emphatic pronouns are present or in the perfect tenses.

Auxiliary verbs in Alashian are secondary verbs that modify the main verb in some way, typically introducing new modality information. Syntactically they are notable for carrying all tense information instead of the main verb, and forcing the main verb to appear in the perfective subjunctive.

19.2 The Indicative Mood

19.2.1 The Present Tense

The present tense is used to denote actions taking place at the present moment:

Habitual actions that have taken place before and are expected to continue taking place typically used the present tense as well, whether or not the action itself is actually taking place at the present moment:

It also is used for actions that started in the past and continue into the present:

As in many other languages, the present tense may also describe events that are to take place in the near future; the timing is typically implied by adverbs such as “tomorrow” or “soon”.

Within narration, present tense forms may also mark past events. This is usually known as the 'historical present'.

19.2.2 The Preterite Tense

The preterite tense marks actions that were completed in the past.

This includes actions that took place over a protracted period of time, so long as the action is being treated as a whole and not as a series of repeated events each with their own beginning and end. Compare, for instance, the following two sentences, the first of which requires the preterite, the second the imperfect.

19.2.3 The Imperfect Tense

The imperfect tense marks a past action as a process with an internal temporal structure beyond the simple 'start' and 'end' model suggested by the preterite tense. It is thus used for a number of different types of actions.

The imperfect is used when one action takes place during or interrupts another action, since such an interruption indicates the existence of a temporal 'middle'. The interrupted action takes the imperfect, while the interrupting action takes the preterite.

The imperfect is used when an action is habitual or repeated, since this indicates the existence of many start and end points.

If the duration of an action is specified, it will generally appear in the imperfect, since duration implies process. However, this does not hold if the verb has an implied perfective aspect. Contrast the two sentences below, where the verb appears in the imperfect in the first case and preterite in the second. This is because “read” is contextually perfective in the second example, and it could be thought of as an instantaneous action of “finishing reading” taking place after a period fifteen minutes, rather than a prolonged action of “reading” taking place over fifteen minutes.

Another pair, the first using the preterite, the second the imperfect:

19.2.4 The Future Tense

The future tense is a complex tense formed using the auxiliary verb *lək plus the perfective subjunctive. It marks any sort of future action, whether structurally perfective or imperfective, once or repeatedly.

19.2.5 The Present Perfect Tense

The present perfect is another periphrastic tense, although with slightly different syntactic behavior. The subject is marked by a mandatory genitive pronoun or genitive construction with the preposition λι- li-, with nominative case pronouns and clitics forbidden, while the main verb appears in the perfective subjunctive. The adverb πλέ ple “then” is also nearly always present in the same clause.

The present perfect marks a past event with a result that has present relevance. In this sense it is similar to the English perfect, but with the added condition that the resulting state must still be true at the present time. Thus, a sentence such as “I have opened the window” (present perfect) implies the window is still opened, while “I opened the window” (preterite) makes no statement as to whether the window was subsequently closed. In most cases the present perfect can be replaced by another tense and still be fully grammatical, just with a slightly different emphasis on consequences.

The present perfect is also frequently used to mark reported information that the speaker cannot personally attest to. It is thus quite commonly used in clauses introduced by verbs of communication.

19.2.6 The Pluperfect Tense

The pluperfect tense (or past perfect) is quite similar to the present perfect; the only syntactic difference is the inclusion of an invariant verb υή wē “it was” immediately before the verbal complex. This form has the same purpose as the present perfect, simply with a past reference point rather than a present one. In other words, the pluperfect expresses an action with consequences relevant for a particular point in the past, or indirect evidentiality in reference to a past action.

19.3 The Subjunctive Mood

The subjunctive mood in Alashian has two main purposes: marking the main verb of a clause when accompanied by an auxiliary, and marking the main verb of an irrealis subordinate clause. Unlike in the indicative mood, Alashian has two subjunctive mood forms that are distinguished by aspect, not by tense: a perfective subjunctive and an imperfective subjunctive. These two forms have very different origins, and so demonstrate very different syntactic behavior.

A third form, the volitive, is usually classified among the subjunctive forms due to its formation. It fulfils a number of non-subjunctive irrealis functions, in particular the optative, hortative, and deontic moods.

19.3.1 The Perfective Subjunctive

As can be seen in the formation of complex tenses, the perfective subjunctive is always used when an auxiliary verb is present, with the auxiliary taking indicative forms and the main verb subjunctive forms. This is also true of modal auxiliaries, such as the capacitative в̄άραδ varad “be able to X” and ventive βού bū “X to here, X towards me”. However, this does not apply to all two-verb constructs; verbs such as χάσαρ xasar “want [to X]” and μανώ manā “forbid [from Xing]” are not considered auxiliaries, and so require infinitives rather than subjunctive mood verbs. The distinction between auxiliary verbs and secondary verbs is elaborated upon in section 19.8.

Note that no coordination is needed when an auxiliary verb is present; the subjunctive verb is simply placed after the indicative auxiliary. Adverbs may intervene between the two verbs, however.

The perfective subjunctive also makes an appearance in purpositive irrealis clauses, that is, in clauses expressing intended purpose or result. This form is therefore frequently seen after verbs like χάσαρ xasar “want” or ναββήτ nəbbēt “hope” ¹ , impersonal adverbs such as ισ̄αλλώ 'išallā “hopefully”, and the conjunction λίκ lik “so that, lest”. Interestingly, in all three situations, the perfective subjunctive verb may appear either with or without the conjunction lik, so that such verbs may immediately follow other verbs serially; in the modern language this is generally viewed as a null conjunction, though historically it is the result of the conjunction ve- “and” merging with an older perfect form (cf. the waw-consecutive in Biblical Hebrew).

In conjunction with verbs in the past tense, the perfect subjunctive may also mark consequence, taking the place of conjunctions such as “[such] that”, “and then”, and “consequently.”

19.3.2 The Imperfective Subjunctive

The imperfective subjunctive is a derivative of the present tense, and so has an additional progressive or imperfective aspect to it. Unlike the perfective subjuntive, it may appear as the sole verb in an independent clause, though it is far more common as a marker of irrealis statives or progressives.

In independent clauses, the imperfective subjunctive may only appear negated, where it marks both present/future tense and some degree of doubt on the part of the speaker. Such clauses can often be translated as “I do not foresee that...”.

It is also employed in counterfactual present- and future-tense statements. The former is most commonly seen in conjunction with the adverb λω...αδ lā...'ad “not yet”, while the latter appears most often with the conjunction τζιήν čien “when, once” in its counterfactual sense.

Finally, the imperfective subjunctive appears in substantive clauses with imperfective meaning; that is, in clauses fulfilling the role of the direct object of a verb such as ιαδώ yadā “know”, ρώ rā “see”, or υρδινώ 'urdinā “command, order”. The conjunction με me is used as a subordinator.

19.3.3 The Volitive Mood

The volitive, when used by itself, indicates wishes, hopes, and desires, and corresponds with English modal particles like 'let' and 'may'.

The volitive is also used to express promises or threats, particularly in the first person.

In conditional sentences, the apodosis (result clause) will appear in the volitive if a) it has future meaning and b) is something the speaker views as beneficial or otherwise in a positive light. This usage likely originates from swearing oaths.

Weak obligation (“should”) can also be expressed by the volitive. Stronger obligation (“ought”, “must”) requires other auxiliary constructions, and so does not need the volitive.

The volitive is also typically used in questions that expect an imperative verb in response, or at least an implied imperative.

19.4 The Imperative Moods

The imperative mood forms mark commands. They are a defective conjugation, existing only in the second person. They also have a system of verbal negation separate from the other verbal forms.

19.4.1 The Imperative Mood

The imperative forms are used for most general commands.

The imperative is negated with the negative marker ελ 'el and the enclitic ακ/'κ (ə)k. However, the presence of another enclitic (i.e., a direct object pronoun) will displace (ə)k.

19.4.2 The Precative Mood

The precative mood is a milder form of imperative used to mark requests and encouragement. Negation works the same way as in the imperative.

19.5 Negation

Most verbs can be negated using the preverbal adverb λω lā.

Imperatives and precatives use the adverb ελ 'el instead (since prohibitory negation is fundamentally very different than factual negation). This is typically accompanied by a post-verbal enclitic k or ək, although this may be dropped if a clitic pronoun attempts to fill the same slot.

In indirect commands (which always appear in the perfective subjunctive), either lā or 'el (without the (ə)k component) may be used. The difference is mostly dialectal, with lā dominating in the north and 'el in the south.

In multiple-verb constructions employing an auxiliary, only the auxiliary may be negated. Double negations (such as “I can't not go”) require rephrasing (“It cannot be that I do not go”). This contrasts with the purpositives in sentences 73 and 74 above by the fact that auxiliaries cannot be followed by the conjunction lik “so that”, while the purpositives may.

In multiple-verb constructions employing an infinitive, either or both verbs may be negated independently of the other. However, infinitives, being nominal in form, cannot use adverbial negation as verbs do; they must be negated either with the pseudo-adjective ήμα 'ēma or the prefix ει- 'ī-, with the former being far more frequent. The same is true of substantive clauses with an infinitive component.

Negative adverbs in the protasis of a conditional sentence typically fuse with the conditional conjunction. In factual conditions, μίρ mir and λω lā fuse to become ιλλώ 'illā “if not” (the form coming from an older conditional 'in “if”), while in counterfactual conditions, λού lū and λω lā merge into λυλλώ lullā “if [it were] not”.

19.6 The Copula в̄εί vī “to be”

The verb в̄εί vī (root *hwāy) means “to be”, and is the main Alashian copula used to link together multiple noun phrases. Its conjugation is highly irregular, as is its behavior.

In the present tense the copula generally surfaces as zero, with the subject and complement simply juxtaposed (the so-called “nominal sentence”). If the subject is a pronoun, it will always appear in its full form, never as a clitic.

The present tense forms can be used for emphatic meaning. Subject pronouns may appear as clitics, but due to the emphatic nature of non-zero copulas in the present tense, full pronouns are far more common.

In other tenses the copula is always overt.

In the preterite and imperfect tenses, the third person forms have two variants: the 'weak' forms, which serve as the normal copula, and the 'strong' forms, which are used as existentials (see section 19.10). Since the act of 'being' is inherently imperfective, the imperfect forms are the most common in the past tense. The preterite forms of 'be' are used only in the context of other preterite verbs to refer to states at a particular point in time; this is particularly common in substantive clauses.

19.7 Voice

Alashian has two voices, the active and passive. These are distinguished morphologically for the most part, so that syntactically there is little difference between a passive verb and an intransitive active verb.

The one major morphological difference is that most passive verbs (other than those in Scale I nuktāb) lack an imperative and precative. Instead, a periphrastic construction consisting of the conjunction lik “so that” plus the second person perfective subjunctive must be used. In this case, there is no formal distinction between the imperative and precative.

As can be seen above, not all morphologically passive verbs are semantically passive, usually due to gradual semantic drift. In addition to νίστυσ̄φαν nistušfan “admit” above, some other such 'deponent' verbs include: εννυτώρ 'ennutār “remain”, νίστυσαβ nistusab “happen”, νυμτώρ numtār “rain”, νίστυσκαβ nistuskab “surrender”, and many others.

The agent of a passive verb may be marked with the preposition ηαλαδεί haladī “by”.

The one time passive verbs are able to take a direct object is in the 'internal object' construction, where an infinitive cognate to the main verb is used to indicate emphasis or totality. Internal objects are discussed further in section 19.13.

Scale V (nitkatab), while normally reflexive, can also serve as a mediopassive. Morphologically and syntactically it patterns as active, however.

19.8 Secondary Verbs

Alashian secondary verb constructions involve a finite main verb plus an infinitive representing the secondary verb. The main verb must be transitive; secondary verb constructions are, in effect, a means of replacing the direct object of the main verb with another verb, as can be seen in the examples below.

However, while these verbs are capable of taking a direct object (as in sentences 100 and 102 above), the infinitive in a secondary verbal construction typically is not syntactically the direct object of the main verb, although it may be semantically. More often than not, the infinitive must be preceded by a preposition; which preposition is determined by the semantics of the main verb. The following possibilities cover the vast majority of cases:

• βι- bi- “by” is used when the secondary verb clarifies the method of performing the primary verb, as in “help X”, “agree to X”, “fail to X” (this last one requires a negated infinitive). In all of these cases the secondary verb actually takes place: helping someone carry something involves carrying, agreeing to speak involves speaking, and failing to complete something (lit. 'not complete') involves not completing it.
• λι- li- “of”, by far the most common preposition, is used when the secondary verb does not represent an action that is actually taking place, as in “want to X”, “like X”, “hate X”, “feel like X”.
• Ø (i.e., no preposition) is required of a handful of common verbs whose semantics do not appear to be clearly differentiated from those requiring li-, as in “remember to X”, “try to X”.

The infinitive may take its own direct objects. Since the infinitives are technically nouns, this is expressed by means of a nominal construct, with the infinitive appearing in the construct state and the direct object immediately following it in the absolute, definite, or partitive state as appropriate. The construct state of all infinitives is identical to the absolute state, even if its surface form would suggest an explicit construct marker.

If the direct object is a pronoun, it may be marked either by possessive suffixes or genitive pronouns following the infinitive.

If the infinitive does not have its own direct object (i.e., is not in a construct), it may appear either in the absolute or determinate states. The determinate state is used when the action is somehow specified to refer to a specific timing or manner by means of adverbs; otherwise the absolute state is the default form. When the infinitive has a direct object, this is not an issue, since the infinitive always appears in the construct state.

Single adverbs may intervene between the main verb and secondary verb, but generally no more than one word can appear in this position.

19.9 Modal Auxiliary Verbs

The modal auxiliaries are a closed set of verbs that are used in conjunction with another verb in the perfective subjunctive to convey some particular modal or aspectual distinction. Verbs in auxiliary position have generally undergone some degree of semantic bleaching; in several cases a single verb may have both a full non-auxiliary form and a bleached auxiliary form, distinguished only by whether it appears in auxiliary position or not. The following overview covers some of the most common auxiliaries.

Alashian has only one defective auxiliary, the future tense marker ιλακ yilək, which only has present tense forms. This is not particularly surprising, however, given its function.

The auxiliary в̄άραδ varad (root *wrīd) indicates possibility, and so is translated “be able” or “can”.

Β̄άκαλ vakal (root *wkāl) indicates permission, translated as “may” or “has permission to”. It can also mark weak instructions.

Ρατζζή rəčhē (root *rčhīy) indicates willingness to perform an action. When not being used as an auxiliary, this same verb means “like, enjoy”.

The verb βού bū (root *bū'), the same as the non-auxiliary verb meaning “come”, has a few different functions. Most generally, it indicates that an action is occurring in the direction of the speaker (a “ventive”), and so is often translated simply as the adverb “here”. In the first person alone, it may also serve as an autobenefactive, indicating than an action was done for one's own benefit. It can also stand in place of a purpose clause to indicate that an action was done with some sort of future utility in mind.

The verbs νατταλή nəthalē “start” (root *hlīy) and less frequently ταννήν tənnēn “stop” (root *tann) often function as auxiliaries, although they needn't necessarily; they may also appear in standard primary/secondary verb constructions with little difference in meaning. Generally speaking, they will be used as auxiliaries if the beginning/end of an action is viewed as distinct from the action above; this contrast is demonstrated in sentences 125 and 126 below.

Multiple auxiliary verbs may be combined. As expected, only the first auxiliary independently shows tense information, while any subsequent auxiliaries are subordinate to it and so must appear in the perfective subjunctive.

19.10 Existentials

Alashian has two types of existentials (“there is/are”): a 'pseudoverb' είθ 'īṯ used only in the present tense and a specialized use of в̄εί vī “be” in all other tenses. Syntactically these two constructions behave slightly differently, given their different origins.

19.10.1 In the Present Tense

In the present tense, existence is expressed using the pseudoverb είθ 'īṯ “there is/are”, a frozen form of an older Semitic root *yṯw "be present" that is now otherwise defunct in Alashian. Είθ 'īṯ in this sense is invariable and typically lives at the beginning of a clause. If in an independent clause, it will typically be preceded by the expletive δε de (which reduces to δ' d before a vowel, as with είθ 'īṯ); in dependent clauses, no such deictic element is used.

Είθ 'īṯ can also mean “be present” in all persons, a relic of the particle's origins. When the thing that is present is an actual noun, it behaves identically to είθ 'īṯ in its existential sense. If the thing that is present is a pronoun, however, then accusative clitics are used, placed before είθ 'īṯ in independent clauses (displacing de) and after είθ 'īṯ in dependent clauses.

The negation of είθ 'īṯ is λήθ lēṯ “there isn't/aren't”, a contraction of lā 'īṯ. It behaves identically to είθ 'īṯ.

19.10.2 In Other Tenses

In other tenses, existence is expressed using third person forms of в̄εί vī “to be”, conjugated according to tense and agreeing in number with whatever is being marked for existence. Unusually for в̄εί vī, the verb typically appears at the very beginning of the clause, and in independent clauses it is usually accompanied by a nominative clitic pronoun. In the past tenses, the strong (non-reduced) third person forms are always used, never the weak (reduced) forms.

Negation is handled regularly, using λω lā or ελ 'el as appropriate.

The “be present” meaning can also be expressed with в̄εί vī used in this way. If the thing that was/will be present is a personal pronoun, the third person forms of в̄εί vī (again, agreeing in number) will be accompanied by an accusative clitic pronoun with the correct person marking. This unusual agreement pattern is likely caused by influence from the present tense forms.

19.10.3 In Possessive Constructions

Straddling the boundary between Semitic- and Indo-European-speaking territory, Alashian has picked up both Semitic 'locative-type' and Indo-European 'have-type' possessive constructions, the former being inherited and the latter being developed under Greek influence.

'Locative-type' possession requires the use of an existential in conjunction with the preposition of possession λι- li- “of”, so that a sentence such as “I have a book” is expressed as “There is a book of me”. This is the standard construction used in most cases of possession whenever the thing possessed is a tangible object.

The negation of a locative-type possessive construction simply entails negating the existential component.

'Have-type' possession involves the use of a dedicated verb, specficically ləkhař (root *lkhāř) “have”. It is generally used whenever the possessed object is abstract, as well as in many idiomatic constructs; for instance, it is commonly used with references to food or drink to mean “partake in” (cf. English “have breakfast”, “have a drink”, etc.).

19.11 Pronominal Clitics

Alashian has two sets of pronominal clitics, one representing the nominative series of personal pronouns and the other the accusative series. These clitic pronouns are ubiquitous in Alashian; the majority of verbs are accompanied by them, though the rules for when they are used and where they are positioned are somewhat involved.

It is first necessary to distinguish how clitics are used with verbs in independent clauses versus subordinate clauses.

19.11.1 In Independent Clauses

In independent clauses, clitics always precede the verb in all forms except in the imperative and precative moods. This normally takes the form of a nominative clitic; only if the direct object of the verb is a pronoun can an accusative clitic displace the nominative one. Only one clitic may be used with a single verb, so if an accusative clitic is used, the nominative clitic will drop entirely.

If there is no direct object, the nominative clitic will always be present, even if the subject of the verb is overt or if it is an emphatic pronoun. This is not true of accusative clitics; if the direct object is a noun phrase, then no accusative clitic can be used.

In the imperative and precative, nominative clitics may never be used. Accusative clitics may be used if the direct object is a pronoun, but in this case they always follow the verb, never precede it.

If the verb is negated, nominative clitics become optional; if they are used, they come before the negation particle. Accusative clitics may continue to be used, but they always come after the verb.

If the direct object of the verb is an animate noun phrase marked with τα ta (discussed in the next chapter), accusative clitics will be used instead of nominative clitics, the only time the accusative is allowed to be doubly marked.

19.11.2 In Subordinate Clauses

In subordinate clauses nominative clitics may not be used at all. In situations where a nominative clitic would be used in an independent clause, the verb will appear unaccompanied by pronominal clitics in subordinate clauses.

Accusative clitics will always follow the verb.

19.11.3 In Complex Verbal Constructs

Nominative clitics in two-verb constructs are regular, always appearing in front of the entire verbal construct.

Accusative clitics are not used in secondary verb constructions; these indicate direct objects using genitive formations.

In auxiliary verb constructions, accusative clitics are either placed before the auxiliary verb or after the main verb, according to the normal rules.

Perfect verbs are more complicated, given their syntactically impersonal nature. Nominative clitics are never used with perfects to mark subjects, but direct object marking is more complex. In standard Alashian direct objects are marked an accusative clitic that is typically placed before the verb (but after the genitive-marked subject), though this moves after the verb if a negative particle is present. However, in some dialects as well as in older texts, the direct object is instead marked with nominative clitics, a trait also found in some idioms even in the standard language. Another common dialectal variation involves always using postverbal clitics, even in positive sentences. In general, the use of clitics with perfect verbs is highly variable across dialects.

19.12 Valency

Verbs in Alashian can take between zero and three arguments.

Impersonal (avalent) verbs in Alashian consist of two types: structurally-impersonal verbs and impersonal passives. The structurally-impersonal group consists of verbs that are inherently incapable of taking any arguments, such as various weather phenomena, (“It is raining”, “It is snowing”), statements of time (“It is Monday”, “It is 9:00”), and headless adjectives or adverbs (“It is cold”, “It is hot”, “It is impossible [that]” “Hopefully”). Impersonal passives refer the use of passive verbs with no explicit subject to imply a general agent (“It is thought that...”, “It is asked that...”, “It is disputed that...”).

All impersonal verbs in independent clauses must be preceded by the syntactic expletive δε de (or δ' d before a vowel), which serves as a dummy subject and occupies the position that a nominative clitic normally would. As with pronominal clitics, however, in subordinate clauses this expletive will not be present.

Intransitive (monovalent) verbs consist of intransitive active verbs, passive verbs, and reflexives/reciprocals. These verbs have a subject, but are incapable of taking a direct object (even if, as in the case of reflexives and reciprocals, there is clearly a logical patient). Transitivity in Alashian is a structural feature; it is inherent in individual verbs, and a transitive verb generally cannot be made intransitive or vice versa without first modifying the verb's morphology. Notice below, for instance, how the transitive and intransitive senses of the English verb “break” are expressed using two different verbs in Alashian.

Transitive (bivalent) verbs are generally active-voice verbs that require both a subject and direct object. There are also a sizable number of 'pseudo-transitive' verbs that require a subject and a prepositional phrase governed by a particular preposition inherent to the verb, typically one of the clitics βι- bi- “by, with”, λι- li- “of, for”, or χι- xi- “to”. Pseudo-transitive verbs have much in common with true transitives, and often the prepositional phrase may be replaced by an accusative pronoun, suggesting that it is in fact perceived as a direct object.

Trivalent verbs take a subject, direct object, and indirect object, the last of which is introduced by one of the same prepositions as above. This group includes a number of primitive verbs mostly dealing with social interaction (“give”, “say”) as well as causatives derived from transitive bases.

Valency switching is typically done morphologically, but not always. For instance, a number of transitive verbs (sometimes referred to as 'ambitransitive') can be used intransitively simply by not expressing a direct object (see “read” below); this is sometimes simply considered an implied object, however.

Passive verbs, which are usually all monovalent, can also be made transitive in one particular construction: the so-called 'internal object', where a verb is paired with a cognate verbal noun (see next section).

19.13 Non-Finite Forms

Alashian verbs have two types of non-finite forms: the participles (or verbal adjectives) and infinitives (or verbal nouns).

All verbs have at least one participle, with katab (Active Scale I) having two. In katab, the two participles have present active and past passive meaning, while the passive participle of nuktāb always has present passive meaning. In other active scales, the participle has present active meaning, while in other passive scales, the semantics are a little more complicated: the participle can have either present passive or past passive senses, generally depending on the semantics of the verb in question and on context, although a past passive sense tends to be more common.

Alashian participles behave just like any other adjectives. They typically cannot take their own arguments, so complex participial phrases do not exist in Alashian; relative clauses must be used instead.

Alashian infinitives, similarly, are true nouns; they cannot have subjects or direct objects except in the form of genitival relationships, and can appear in the absolute, determinate, and construct states (though the absolute and construct states are always identical in form). As abstract mass nouns, however, they have no number contrast. They may serve as both the subject and direct object of other verbs, although in direct object position they will often have to be preceded by various lexically-determined prepositions (see section 19.8 above).

Infinitives may be modified by adjectives, which are often translated to English as adverbs.

Infinitives are also employed in the internal object construction, where the infinitive of a verb is used as the direct object of the same verb, always without the intervention of prepositions, in order to make the meaning of the verb more emphatic. This can be done in theory with any verb, even ones that are normally intransitive. It is, however, somewhat of a marked construction that is not particularly common in colloquial speech outside of idioms.