Double-transitive constructions

Agent-increasing affixes:

Some valency-increasing affixes have a meaning, that itself describes a divalent action; i.e. it contains both a logical Agent and Patient role. Call these Agent1 and Patient1. Some examples of these affixes are:

These affixes are traditionally called double-transitive affixes, although we here also refer to them as agent-increasing affixes, because it seems that the effect of adding such an affix to a divalent stem is to add a new Agent role to the relation. The full list of these affixes can be found through this tag: Double-transitive affix.

The meaning of adding such an affix depends on the valency of the underlying stem. There are two cases:

These roles of the underlying stem are equated with the roles of the affix as follows:

Example 1:

Consider the following examples with V{tqu}V, 'bid': Depending on whether the underlying stem is monovalent or divalent, the meaning thus becomes as follows:

With monovalent stems, the meaning is thus quite straightforward; e.g. from iserpunga, 'I enter', we can make iseqquaanga, 'he bade me enter'. The ending is V{vaaŋa}, i.e. indicative 3sg/1sg. However, with a divalent stem, it becomes more complicated. Consider the following sentences:

This is a double-transitive construction: Evap qitsuk paaraa means 'Eva looks after the cat', but by adding -qqu- to the stem {paarə}V, we introduce a new Subject (mother, who is doing the bidding), whilst the Object (the cat) remains the same. The former subject, bound to the underlying Agent of the stem (Eva), can the optionally be specified in the allative case (here Evamut).

Example 2:

Consider the following examples with V{tət}V, 'let':

In the latter example, the former Agent of the stem, {taku}V, becomes an underlying Agent of the resulting stem, denoted by (someone) in the translation. It can then be specified explicitly in the allative case: Suppose for example we had the sentence Piitap illu takuaa, 'Piitaq saw the house'. By adding V{tət}V, we can transform this into the new sentence

where 'he' is some new, unnamed subject. The house, illu remains the object, but the former subject, Piitaq has now changed role.

See also the other agent-increasing affixes here for more examples of the double-transitive construction.

Valency-reduced stems:

There is one further complication: the underlying stem could be a divalent stem that has had its valency reduced before the double-transitive affix is added. This can happen in four different ways, although only two of them seem to be used in practice:

The first possibility is that the valency could have been reduced by a valency-reducing affix, especially the passive affix V{nəqaq}V, or possibly even a HTR-morpheme. In this case, whichever role remains, whether it be Agent2 or Patient2, will be equated with Patient1 of the affix.

The second possibility arises if the underlying stem is agentive (i.e. non-patient-preserving). In that case, the stem can also drop its Patient role (i.e. Patient2) before the present affix, without any explicit marking (and as with other anti-passive constructions, this omitted Patient role can then optionally be referenced in the instrumental (INST) case). This introduces an ambiguity, where the same resulting word can be read in two different ways. Consider the following examples with the agentive stem {atuvaq}V, 'read', i.e. atuarpaa, 'he reads it' and atuarpoq, 'he reads (something)':

In the first example, the stem preserves its Patient2 role, which is equated with the Patient1 role of the affix and bound as the Object in the ending, whilst the underlying Agent2 role of the stem is shadowed and not bound in the ending. In the second example, the stem silently drops its Patient2 role, and instead it is the underlying Agent2 role, which is equated with the Patient1 role of the affix and bound by the Object in the ending. The ambiguity can be resolved by explicitly referencing the omitted role with a noun in the allative resp. instrumental case:

The third, hypothetical possibility arises if the underlying stem is patientive (i.e. non-agent-preserving). This is, in a sense, the converse of the previous case: Here, the stem can optionally drop its Agent role (i.e. Agent2) before the present affix, without any explicit marking (and as with other passive constructions, the omitted Agent role can in principle be referenced in the ablative (ABL) case). The remaining role is then the Patient role, i.e. Agent1, which then is bound to the Patient role of the affix, i.e. Patient1=Patient2, but notice that this is equivalent to the standard binding of roles induced by double-transitive affixes. Thus, this option does not seem to be used in practice, but it illustrates why adding such an affix often seems to give a passive meaning to the underlying stem.

The fourth possibility is that the underlying stem is reflexive (i.e. both-preserving). In this case, it is possible, that the stem could equate its roles, i.e. Agent2=Patient2, as if the stem had been used with an intransitive ending, and this combined role would then be bound to the Patient1 role of the affix. This construction seems at least theoretically possible. Consider the following example with V{surə}V, 'think that':

Note that this pronoun is itself in the allative case, and since it emphasises the Subject role, it also allows a second reading:

Due to this ambiguity, this construction is likely rarely, if ever, used.

Intransitive usage:

Because these affixes have their own, inherent roles, they also have their own diathesis; i.e. preference for how the roles are bound and equated, if the affix is used with intransitive endings. Specifically, they are reflexive (i.e. both-preserving), so the Agent1 and Patient1 roles will (generally) be equated and bound to the Subject, if one of these affixes is used intransitively without a HTR-morpheme. Consider this example with V{surə}V, 'think that':

In this example, the underlying stem is monovalent, but if the underlying stem instead is divalent we get the same complications as above, with three different possibilities depending on the diathesis of the underlying stem. We shall not go through all of them again, but if the underlying stem is agentive (non-patient-preserving), it can behave like a monovalent stem, so this is straightforward. However, if the underlying stem is patient-preserving (i.e. either reflexive or patientive), we normally get a reflexive double-transitive construction. Consider the following example with V{tət}V:

Thus, the meaning can be viewed as either reflexive or passive, depending on whether one considers the meaning of the underlying stem (where the Agent is removed, and the Patient is promoted, as in passive constructions), or the meaning of the affix (where the Agent and Patient roles are equated and bound to teh Subject, as in reflexive constructions).