The constituent word order of Iridian sentences is SOV, but the agglutinative nature of the language and the presence of case-marking on nouns makes word order typically flexible, with the only universal rule being that the main verb should appear at the end of a sentence.

Topic-predicate constructions

The constituent word order of Iridian sentences is SOV, but the agglutinative nature of the language and the presence of case-marking on nouns makes word order typically flexible, with the only universal rule being that the main verb should appear at the end of a sentence.

The Iridian sentence can be divided primarily into a topic part and a predicate or comment part. The topic is what the sentence is about, while the predicate or comment represents the information presented in the sentence about the topic. While both the topic and the predicate are pragmatic constructs, the topic-predicate construction is important as it determines how the rest of the sentence is structured.

The topic of the sentence does not necessarily coincide with the subject of the sentence. This is true as well in English, as we see in example (1); although where English allows the topic to appear anywhere in the sentence, as long as the subject is placed first, Iridian, typical of topic-prominent languages, requires the topic to always be introduced first, leaving the rest of the information afterwards.

More importantly, the topic of the sentence determines how the main verb, and thus all the other constituents of the sentence, are marked.

As Kiss (2004: 9) notes:

We tend to describe events from a human perspective, as statements about their human participants – and subjects are more often [+human] than objects are. In the case of verbs with a [–human] subject and a [+human] accusative or oblique complement, the most common permutation is that in which the accusative or oblique complement occupies the topic position [.] When the possessor is the only human involved in an action or state, the possessor is usually topicalized[.]

Coordination

Iridian has three groups of coordinating conjunctions: the additive a, ‘and’ and še, ‘with’; the contrastive má and ozná (both translated to English as ‘but’); and the disjunctive/correlative je, le and ni. A corresponds to the English ‘and.’ When coordinating simple noun pairs, however, še, ‘i’s more often used though. The derived construction a še is also common and has a similar meaning to the English ‘and also’.

In constructions with še where one of the nouns coordinated is a pronoun or a deictic, the pronoun or deictic is presented first followed by the other noun in the instrumental case.

In a few cases, a is used instead of še where the latter can be interpreted as having an attributive meaning. Where the noun is marked, however, only a can be used.

Apposition

Appositive constructions in Iridian involve the juxtaposition of two or more noun phrases that have a single referent. An apposition can be non-restrictive if the appositive can be removed freely without changing the meaning of a sentence, or restrictive otherwise.

Formally both non-restrictive and restrictive appositives are treated as modifier phrases but only the latter is grammaticalized. The restrictive appositive must always precede the noun phrase it modifies, linked together by the particle ko. Non-restrictive appositives on the other hand are simply juxtaposed together, although a comma is often inserted around the appositive if it consists of more than one word.

Examples(28a) and (28b) shows two different translations of the English phrase ‘My brother Óto is studying in Munich.’ Example (28a) is nonrestrictive and can be interpreted as ‘I have a brother namsed Óto who is studying in Munich’ while (28b) being restrictive can be translated more on the lines of ‘Among my brothers, it is Óto who is studying in Munich.’ The restrictive appositive implies specificity and by extension the existence of a group where this specificity holds true; in (28b) this is taken to mean that a set of brothers exists and Óto is a member of this set.

The English appositive oblique, which uses an ‘of’-phrase with an appositive meaning, is translated in Iridian without the particle ko , even though it is restrictive in meaning.

Syntax of event and participant nominals

As we have established in § 3.9.2, Iridian has three forms of nominalisation: (1) the mainly non-productive usage of the nominalising -ou with the verbal stem to form resultant nominals; (2) the use of -ou in conjunction with the gerund-forming prefix po(d)- to form a verbal noun (which we call an event nominal or gerund) and which may either include the internal arguments of the parent verb or not; and (3) the formation of a participant nominal (cf. Pearson 2013) which nominalises not the event described by the verb but its participants.

Since gerunds represent the nominalisation of the event described by the verb, they are therefore inherently abstract and active in meaning. Since the nominalised forms are abstract, it follows that they are also tenseless and aspectless. Iridian gerunds, however, may be optionally marked for their lexical aspect or aktionsart using the continuous aspect suffix -eví (which subsequently becomes -ív- through sound change). It is important to note though that although a marker for grammatical aspect is used, what is being marked is lexical and not grammatical aspect; specifically, the addition of -ív- only signifies that the action is iterative in nature and thus the gerund itself remains tenseless and aspectless.

In CENs, both the agent and the patient are marked in the genitive.Šereikaitė (2020) argues that although (in the case of Lithuanian, at least) the actor and the theme from the original sentence both become marked in the genitive in the resulting complex event nominal, the superficially indentical genitives are actually two distinct cases: a higher genitive (gen.h) assigned to agents and possessors and a lower genitive (gen.l) assigned to grammatical objects. Although this argument is interesting and probably holds true as well in Iridian CENs, we will not make an effort to ascertain whether there is an actual difference in the two genitive cases in Iridian as this is not needed for the purpose of this grammar. If both are present, the agent must always appear first. This construction is quite common cross-linguistically, as we see in the examples below.

The use of the genitive to mark both the actor and the theme in the original sentence is of course a recipe for ambiguity. When only one of either the actor or the theme is present in the CEN, the ambiguity is on whether the noun marked represents the one or the other, as, e.g., the phrase Jancě podohletou which can be interpreted to mean either ‘the act of remembering Janek’ or ‘Janek’s act of remembering’ without any further information. A second ambiguity arises when both the actor and the theme are in the sentence as it is unclear, without any context, the genitive is actually being used to mark their thematic role in the originally or is in fact a possessive. The same is true in, for example, Lithuanian where as Šereikaitė (2020) points out, sentence (30b) can also be alternatively translated as ‘[the] destruction of Jonas’s plants’.

The first type of ambiguity is resolved in English by using word order: in general, a prepositive genitive (i.e., using the clitic ’s or the possessive form of a pronoun) is used when the noun in the genitive case in the CEN represents the actor (e.g., ‘John’s remembering’) while a postpositive genitive is used when the noun in the genitive represents the theme (e.g., ‘the remembering of John’). This in turn, can be extended to the second type, e.g., ‘John’s remembering of Margaret’. However, the obligatorily head-final nature of Iridian syntax means that such strategy is not possible. Instead, the strategy used in Iridian is more similar to the one found in Tagalog where the theme may be marked using the oblique saThis becomes kay before proper nouns instead of the genitive ng.To call ng (pronounced [nɐŋ]) as a genitive marker is simplistic (even erroneous) but should be enough for the purpose of our discussion.Thus we can restate (30c) as follows:

An immediate consequence of replacing the genitive ng with the oblique marker sa/kay is that the theme is now interpreted as definite (cf. Kaufman 2009: 3, 40). The use of the oblique to mark the theme can be used even when only one element is present in the event nominal; in fact, when the theme is known as definite for a fact (e.g., if it is a person), the choice between the oblique and the genitive is what distinguishes the actor and the theme. Thus we have

In Iridian, the a na clause corresponds to the Tagalog use of the oblique to indicate a definite theme in a CEN.

Quotative constructions and evidentiality

Quotative construction in general

Evidentiality is a grammatical category that is concered in the explicit encoding of a source of information or knowledge (i.e. evidence) which the speaker claims to have made use of for producing the primary proposition of the utterance (Diewald and Smirnova 2010: 1-2). Iridian is unique among languages of Central Europe (and of Europe in general) in possessing a grammaticalised evidentiality system. Even non-Indo European languages in the region such as Hungarian (cf. author) or Basque (cf. Alcázar 2010) do not possess an overt evidential. Of course a speaker’s source of information may be expressed through other methods The Iridian evidentiality system more or less falls under Aikhenvald’s (2004) A3 category, where the distinction is between the marked quotative form for reported speech/hearsay and the unmarked ‘everything else’ category which is evidentiality-neutral.

Quotative constructions and reported speech

The principal use of the quotative is to explicitly mark reported speech. The reported clause is separated from the rest of the sentence by the particleto-že.

The particle to-že is in fact made up of two separate clitics: the particle to which is used to mark relative clauses, and -že which is the primary quotative particle. This is made more evident in nested quotations, where -že can only be attached to the rightmost reported clause:

The use of pronouns in quoted clauses is similar to English, with the main exception being the use of the reflexive se if the subject of the quoted clause is the same as the subject of the main clause. This is true even if the subject of the main clause is a pronoun.

The verb zěká, ‘to say’ is called a verbum dicendi from the Latin meaning ‘verb of speech/speaking.’ Other verba dicendi in Iridian include vadá, ‘to think’; kvuštá, ‘to hear’; vidá, ‘to see’; hloupá, ‘to ask’; ohletá, ‘to remember’; sehová, ‘to recount, to tell a story’. Note that although they are called verbs “of speaking” they do not necessarily introduce speech as much as function as grammaticalized tags marking the quotative, which is more properly analyzed to mark not just speech but inferentiality and evidentiality as well.

More complex verba dicendi can be formed by using an imperfect converbial construction (the converb form in -ěc) with a canonical verbum dicendi. To illustrate this consider the following sentences in English:

We see that both said (46a) andwhispered (46b) are verba dicendi in English. Nonetheless it’s also obvious how 46b is simply a function of (46a), i.e., we can express (46b) in terms of (46a), in this case using an adverbial construction (‘in a whisper’) as we see in 46c or the more affected 46d. Finally using a simple adverbial is theoretically allowed in English (46e), although as we see the resulting construction is rather unwieldy or unnatural-sounding.

In Iridian, however, constructions like (46b) are not permitted, with preference given to adverbial (or more correctly, converbial) constructions. Thus we translate (46b) as:

It should be noted as well how the verb vadá, ‘to think’ and its derived forms, due to their inherent meanings, require the subjunctive to be used in the reported clause. This is true whether or not the subjunctive would have been used had the reported clause been a regular dependent clause.

We see from that when it comes to reported speech and similar constructions in Iridian, the verbum dicendi is not necessary to create a well-formed sentence. The same is true with the quotative particleto-že. Both can be omitted without making the sentence grammatically incorrect since the quotative particle is enough to identify the reported clause.

In most instances, however, removing either the main verb or the main verb and the quotative particle can cause the resulting sentence to acquire a new meaning. This is especially true when the quotative mood is used not to report speech but to imply a certain unsureness on the part of the speaker about the information being presented, or for the speaker to distance themself by implying through the use of the quotative that the information is secondhand and not theirs. Generally to-že is kept when the speaker is quoting themself, to repeat or emphasize what they have said, or expletively, to express their frustration or affirmation.

Interestingly, commands and requests are not treated as reported speech but as regular subordinate clauses governed by to and not by to-že.

When the quoted clause is a question, whether a direct one or not, the quoted clause is preceded by the particle a, ‘and’ and the word ane, ‘whether’ is used instead of to-že. The word ane is also used for verba dicendi that are interrogative in nature, such as préhoustá, ‘to ask’,

Bare quotatives and clause linking

Quoted clauses in Iridian may also appear without an overt predicate, as well as without being signalled by the quotative particle to-že. We will call this construction a bare quotative after the terminology in Tomioka and Kim (2019) in reference to embedded quotative constructions in Japanese and Korean without overt predicates. The term as originally used by these authors refer only to embedded quotatives in Japanese and Korean, but we will be using it to refer to both an unselected (i.e., predicateless) quotative in a subordinate clause (which we will call syntactic) and in the main clause (which we will call semantic).

The choice to call the second type a semantic bare quotative is motivated by the fact that an unselected quotative in the main clause is often used not to mark a speech act but to indicate the epistemic value of (viz., to pass the speaker’s judgement on) a proposition. Nevertheless, we can still see it used as a true quotative, as when the omission of the predicate or the quotative particle is through mere ellipsis.

The first type, on the other hand, is mostly used as a clause-linking strategy. The quotative construction is still considered as a speech act, but, like converbial constructions or še clauses, the relationship between the main clause and the reported clause becomes interpreted as being one of causality, or at least of dependency, although of course this causality or dependency is only indirect, as we see in the examples below, where the embedded quotative and the simple še clause present to different interpretations.

Epistemic extensions

As in most other languages with an overt evidential system, the Iridian quotative has secondary epistemic extensions. This may be realised either by using the quotative by itself or through auxiliary epistemic markers. As we have established in the previous sections, the quotative can be used by a speaker both to distance themself from the statement on the one hand and to assert their belief in its truthfulness on the other; the use of a secondary epistemic marker eliminates this possible confusion in what would otherwise have been a contradictory usage of the same grammatical category. These auxiliary particles nonetheless, may of course be left out in discourse if the speaker thinks the epistemic usage of the quotative is clear enough from the context.

A speaker’s judgement of the truthfulness of a statement may be made clear by the dubitative bude or the affirmative toleto. When using the quotative to quote oneself, bude expresses a disbelief predicated upon surprise rather than on a judgement of a statement’s veracity; used the same way, toleto acquires a secondary meaning of insistence, even annoyance.

A speaker’s uncertainty may also be expressed using the quotative even when the statement directly came from the speaker. The uncertainty may refer to both the factuality of the statement or to its source. This strategy is used to signal the speaker’s emotional or cognitive distance from the event. This may be further complemented by the particle iz which we will glossing here as rep for reportative but only for the sake of convenience, in order to distinguish the various auxiliary particles we have introduced here, as the “reportative” does not exist as a true grammatical category in Iridian for our purposes. Iz implies a greater degree of disjunction between the speaker and the statement than the plain quotative. Although it does not pass a judgement on the truth value of the statement as do dube or toleto, iz makes it clear that the statement did not come from the speaker and that the responsibility for the statement does not lie on them. Iz is particularly common in newscasts or in other formal settings where the speaker is communicating statements from another speaker or group and the identity of the speaker or group has already been established earlier in the conversation and is thus known to everyone.

Uncertainty on the truthfulness of the statement may also be expressed using the inferential particles bylo and atole. Whereas iz raises the question of the character of the source and is neutral as to the speaker’s commitment to it (although one can be understood simply by pointing out the fact that the source is something other than oneself to be effectively passing judgement) both bylo and atole reflect the speaker’s judgement. Bylo in general is used when the proposition is coming from the speaker themself while atole is used when the speaker thinks that the statement can be inferred from the surrounding facts.

Comparative constructions

The clitic tám is used to form simple comparative and relative constructions. Tám is often ommitted where the comparison can be implied from context. In this construction, the standard of comparison (the noun preceded by ‘than’ in English) is unmarked and the noun being compared marked in the agentive if it is a positive/negative comparison, or in the instrumental if it is a correlation.

Note that tám can only be used with the copulative form of the stative verb, as the attributive and nominal forms have separate conjugated comparative forms. When using these forms, however, the standard of comparison is marked in the genitive. In relative constructions, the instrumental is also replaced with the genitive, but the modifier zní, ‘same’ is added before the stative verb.

Tám can be relativized by appending the clitic to. When used with tám-to the standard of comparison is marked in the patientive case. The use of tám-to in relative clauses is discussed in further detail in the next chapter.

Questions

Yes-no questions

A declarative sentence can be made into a question by a simple rise in intonation at the end of the phrase:

Yes-no questions, especially longer ones, may also be formed using the clitic no, which immediately follows the element of the sentence being questioned. To question the sentence as a whole, no sentence-initially. No may also appear after other elements of the sentence, but the resulting word order is generally more emphatic and often includes promoting the element where no to the topic position and the nominalisation of the resulting verb phrase if possible.

Content questions

Content questions, also known as wh-questions, are formed using the interrogative pronouns jede, ‘who,’ ježe, ‘what,’ jena, ‘where,’ etc.11 Iridian requires the wh-phrase to be moved to the beginning of the sentence, thus causing it to occupy the topic position. This wh-fronting consequently causes the voice of the main verb to be reframed to accomodate the new topic. More commonly, especially colloquial Iridian, this also means the nominalisation of the main verb phrase, essentially making the question an equational sentence.

Alternatively, the element being questioned may be replaced with a question word without changing the original word order, in which case the addition of the clitic no is required. Note that questions formed this way generally have a more emphatic meaning.

Wh-fronting may sometimes cause peripheral elements of a phrase to be moved together with the wh-item to the beginning of the sentence, a phenomenon linguists call ‘pied-piping’ (Ross 1967: 263-4). When this occurs, Iridian is more conservative than English in that it usually keeps the same question word instead of replacing it with a specialized one (in English, normally, ‘which’); it may, however, use jak, ‘which’ if the expected answer to the question is an element of a class, i.e., not unique. Consider, for example, the two questions below:

Existential construction

In general

An existential sentence is a specialized construction used to express the existence or presence of someone or something. The particle ješ and its inverse niho are used to form existential sentences.

The existential construction in Iridian was originally a locative one, and this could still be seen in how the use of ješ and nihorequires both the noun or noun phrase whose existence is posited and the location where such existence is said to be true to be explicitly present in the sentence. In true existential sentences (e.g., ‘There is a God’ or ‘There is still hope’) where the argument is the existence of something and not just it’s mere presence somewhere, the patientive form of the reflexive verb se, sní, is used. In addition, where this ostensible location is present in the sentence, it would occupy the topic position in the sentence, and unlike in regular sentences, must be explicitly marked in the patientive.

Possession

Existential constructions are also used to indicate possession, with the possessor marked in the patientive case.

Impersonal constructions

Iridian prefers using existential constructions where English and other IndoEuropean languages would use indefinite pronouns. More formally, sentences of this type are called impersonal constructions.14 In general an impersonal construction in Iridian is produced by nominalizing a verb phrase which would otherwsise have been the predicate of an indefinite pronoun. We can illustrate this in English as follows: