Commentary: Use of verb suffix -moH

This page proposes a theory which attempts to explain the behavior of the Type 5 verb suffix -moH. My theory is derived from the canon phrase ghaHvaD quHDaj qawmoH Ha'quj 'his sash reminds him of his heritage.' [Skybox card 20] and is an attempt to understand why and how it works, and in the process to propose that it is univerally applicable to all transitive verbs with –moH.

My theory hinges on the principle of a hierarchy of semantic roles that nouns can play in relation to verbs in the Klingon sentence, in the order agent-patient-indirect object. In Klingon, these roles are usually embodied in the parts of speech (in order) of subject, object and benefactor (which is expressed in Klingon by the Type 5 noun suffix -vaD). All verbs have at least one associated role, the subject. The addition of the suffix -moH adds a new agent to the verb phrase, causing the original agent to move to the next empty role in the phrase.

We seem to have pretty good concensus on what happens when -moH is added to an intransitive verb: the original subject moves to the next unoccupied role in the hierarchy of the verb phrase, to the object. Example: pIl loD 'the man is inspired' > loD pIlmoH be' 'the woman inspires the man (== 'the woman causes the man to be inspired').

But what happens with transitive verbs, which already have filled object roles? (Let me note here that transitive verbs are those which carry the implication of an object to complete their meaning at the lexical level, whether the object is actually present or not.) I assert that our canon example shows that the original subject is again displaced, but this time not to the object role, which is already filled (at least potentially), but to the next empty role, the indirect object, marked in Klingon by the suffix -vaD. Example: tlhIngan Hol ghoj loD 'the man learns Klingon' > loDvaD tlhIngan Hol ghojmoH be' 'the woman teaches the man Klingon'. Since we are taking about roles filled at the lexical level of the verb, there is no change in this construction if the object is not actually present: loDvaD ghojmoH be' 'the woman teaches the man'.

The idea of a hierarchy of roles is part of a linguistic principle called a valence increasing operation. According to this principle, every verb has a natural set of actors and most languages have ways to add more actors to the verb phrase. Adding a causative meaning (in Klingon, by adding the suffix -moH) to an intransitive verb increases the number of participants, and thus the valence, from one to two; adding it to a transitive verb increases valence from two to three.

There is another way to look at the behavior of –moH with transitive verbs, focusing on the idea of causation as a type of clause embedding (an embedded clause is a verb phrase that is contained within another verb phrase). Let’s look at a clause and how it can be embedded in a couple of types of verb phrases, and then how these are realized in specific languages. (I’ll use the formula VERB(agent,patient), as a shorthand to show  the underlying meaning of the phrase, exclusive of the particular way the meaning is expressed.)

Formula:                     LEARN(son,Klingon)

English:                      The son learns Klingon.

Japanese:                  musuko-wa kuringon-go-o manteiru.

                                    son-subj Klingon-obj are-learning.

Klingon:                      tlhIngan Hol ghoj puqloD.

This is the basic sentence, stating a simple fact, using the language rules and resources of each language.  Agent and patient are embodied in the roles of subject and object, expressed in terms of each language (position in English and Klingon, and particles in Japanese).

Formula:                     WANT(father,LEARN(son,Klingon))

English:                      Father wants the son to learn Klingon.

Japanese:                  ofusan-wa musuko-ni kuringon-go-o manatai.

                                    father-subj son-I.O. klingon-obj wants-learn.

Klingon:                      tlhIngan Hol ghoj puqloD neH vav.

This demonstrates that the embedding of a clause in a WANT formula is embodied in different ways in different languages.  In English, the agent of WANT enters the phrase as the subject of “wants”, while the embedded LEARN clause becomes the patient of “wants” by conversion of the embedded verb to an infinitive, with the embedded verb retaining its own agent and patient.  An archaic way of expressing this makes the relationship clearer: “Father wants that the son learns Klingon”.  In Japanese, the idea of WANT is expressed by a suffix –tai on the verb manabu, the new agent takes over the subject role from musuko, the agent of the embedded clause, and musuko is moved to a subsidiary role, outside the VERB(agent,patient) formula.  In Klingon, WANT is expressed, as in English, by its own verb and agent, and the embedded phrase is repeated without change, acting as the patient of neH.  This is made clearer in other phrases that require the pronoun ‘e’, eg.  tlhIngan Hol ghoj puqloD ‘e’ vIHar.

Formula:                     CAUSE(father,LEARN(son,Klingon))

English:                      Father teaches Klingon to the son/teaches the son Klingon.

Japanese:                  ofusan-wa musuko-ni kuringon-go-o oshieru.

                                    father-subj son-I.O. klingon-obj teaches.

Klingon:                      puqloDvaD tlhIngan Hol ghojmoH vav.

Causative phrases employ the same type of embedding.  The new agent, the one causing the action, is often referred to as the causer, and the one compelled to act (the agent of the embedded clause), the causee.  In English and Japanese, we see that the notion of causation is expressed at the lexical level.  That is, the English word expressing the idea “cause to learn” is “teach”, and it is a separate word in Japanese, also.  In Klingon, causation is expressed by the suffix –moH.  In every case, the new agent, the causer, is expressed as the subject of the verb, the patient of the embedded verb clause remains the topic being taught, and the causee, the agent of the embedded clause, moves to a tertiary role outside the VERB(agent,patient) formula.  Interestingly, this role is the dative/indirect object in all three languages (made clearer in English by noting that “teach the son Klingon” is equivalent to “teach Klingon to the son” in the same way that “give me the ball” is equivalent to “give the ball to me”).

Actually, there is no inherent reason why the third role of causee should be cast as the indirect object; the valence principle only speaks of a more peripheral role for the original actor, the agent of the embedded clause. But our canon example ghaHvaD quHDaj qawmoH Ha'quj uses -vaD, so evidently the causee moves into the "benefactor" role in Klingon. Interestingly, this is fairly common in several Terran languages, such as Turkish and Quechua (and Japanese!), and is called a dative-shift, although, of course, there is no reason why this should be relevant to Klingon. But the valence principle does specify that the causer becomes the new subject of the verb, and that the patient, or object, does not change its role.

For that matter, there is (or was, before our canon phrase above) no inherent reason why verb objects with –moH should behave as they do. Before we had the canon phrase, there was much debate about what happened to transitive verbs when –moH was added.  Many believed (and, to be honest, some still do) that, since adding –moH to an intransitive verb moved the original subject to the object position, this would be true of all verbs, transitive verbs included.  In other words, tlhIngan Hol ghoj puqloD would become *puqloD ghojmoH vav, and there would be no role for the original object, tlhIngan Hol, in the new phrase.  On the face of it, this possibility was as likely as the other. This belief was perhaps influenced by the dictionary word ghojmoH, glossed as "to teach".  But this ignores several factors:

a.      Okrand himself said that this particular word was included only to make it easier to find the word in the English-Klingon side;

b.      He has also said that it should be understood as ghoj+moH and not as a unique verb in its own right; and

c.      The English verb "to teach" can take either person or topic as an object: "I teach Spanish" or "I teach my son.", so the dictionary definition is hardly unambiguous.

At any rate, the appearance of the canon sentence ghaHvaD quHDaj qawmoH Ha'quj should have (to my way of thinking) settled the matter of what happens when you add –moH to verbs.  This essay is an attempt to explain why it happens.

Another interesting linguistic principle may shed light on why the subjects of transitive and intransitive verbs behave differently with -moH. In an intransitive verb, the subject doesn't play the same role as the subject of a transitive verb. The subject of a transitive verb directs the action of the verb to some target, the object. The subject of an intransitive verb is really the experiencer of the verb and not an agent (in poS lojmIt, the door is closed, experiences being closed, but it didn't close itself). Thus, the subject of an intransitive verb has more in common with the object of a transitive verb than it does with its subject. In fact, languages called ergative/absolute languages treat the subject of intransitive verbs grammatically like an object of a transitive verb (eg., by giving them the same case markings), and reserve a different grammatical form for the subject of a transitive verb. Adding –moH to the intransitive verb introduces an agent who performs the action on the experiencer. It's easy for the experiencer to slip into the role of object since it's already halfway there semantically and there is no existing object of the original intransitive verb to block it.

Thus, expressed in the formula used above, "The door is closed" = CLOSED(-,door), illuminating the fact that there is no real agent in an intransitive verb, although in English and in Klingon (but not in Basque, et al.), the patient appears in the verb phrase in the role of subject.  When this clause is embedded in the causative formula, CAUSE(father,CLOSED(-,door)), it is easy to see how poS lojmIt becomes lojmIt poSmoH vav:  the patient of the embedded clause now appears as an object of the causative phrase, and since there is no agent of the embedded clause, there is no need to cast it in the indirect object role.

I should note for completeness that there are at least two other canon sentences that use a transitive verb with –moH (eg. HIQoymoH 'Cause me to hear it!' [TKD 4.2.4]), which may contradict our canon example cited here, but these are inconclusive. They don't have stated objects or causees, and so the 'prefix trick' may be obscuring the grammatical pattern being used. That is, is the SoH addressed by HI- the true object of the verb, or the prefix-trick equivalent of SoHvaD? There's no way to tell. Another alternative sometimes discussed, that every verb behaves differently with -moH depending on its lexical meaning, sometimes casting the original subject as a new object, sometimes as a causee with -vaD or other Type 5 suffixes, I consider pointless to consider, mainly because concensus would never be possible, and so the rule becomes as useless as having no rule at all. Were there truely no rule, I would of course have to accept it and regretfully consider –moH to be basically unusable, but (as I'm sure you can guess) I do not believe this to be the case.

Someone countered recently to the effect that surely I didn’t believe that a Klingon actually made the transformation VERB(agent,patient) to embedded CAUSE(agent,VERB(agent,patient)) in their mind when using –moH.  Well, in fact, I do think that at some level this very sort of analysis occurs.  The whole theory of transformational grammar is based on the idea that some structure in the brain takes thoughts and transforms them into utterances.  Linguists commonly consider such operations as valence increasing or decreasing, clause embedding and clefted sentences as transformations from more basis utterances.  Irregular verbs, the tension between look-up and computation in sentence formation that they imply, and the behavior of children over-generalizing rules until they learn differently (eg. “He goed to the store”) show that somewhere in the brain are structures for analyzing and transforming our ideas into speech. It could hardly be otherwise.  There must be some way for speakers to use the resources of their language to create new statements, or else we are trapped in a Darmok-and-Jalad type of situation, where no one can say anything that hasn’t been said before, although even there, something in the brain is starting with an idea and transforming it into the proper phrase to express it.

In conclusion, I believe that the appearance of the canon phrase ghaHvaD quHDaj qawmoH Ha'quj has answered all our questions about the behavior of –moH with any type of verb. I suggest that my theory of –moH explains all canon examples and predicts the behavior of both transitive and intransitive verbs using a coherent principle to explain that behavior, and gives me the confidence to assert that this is truely how the Klingons use –moH and that we can use it in the same way to form canonically correct sentences.