In all known natural languages, every noun
in any given sentence will fill a certain role in the way it relates to the action described
by the verb. Most verb phrases require at least two roles
to be filled, namely the subject or agent, the noun performing the action, and the direct
object or patient, the noun receiving the action, and some verbs may also require an
indirect object or beneficiary of the action. The words in the phrase that fill these mandatory
roles are called arguments, while most other roles will just provide extra optional information
that’s not essential for the phrase to make sense and are instead termed ‘adjuncts’. It’s critically important then for a language
to have a means of making clear what any given word’s role is, otherwise phrases will be
impossible or at least very difficult to interpret. Across all known natural languages, there
are three main ways this role-marking can be done:
The most common is to have a default word order, so the nouns’ roles can be gleaned
from their position in the sentence. About 87% of languages sampled by The World
Atlas of Language Structures have a default word order of some kind, though there is considerable
variation in word order patterns between, and sometimes within, individual languages. However, using word order alone is comparatively
rare as a role-marking strategy, as it’s most often complemented with one of the two
other role-marking mechanisms: About 60% of surveyed languages make use of grammatical
case, wherein each noun is explicitly marked for its role by means of an affix, while around
80% of languages employ verb agreement,where the verb is marked with affixes that agree
with at least one of the arguments in person, number, and/or gender, allowing the listener
to infer who’s doing what to whom based on the verb’s inflection. Some languages only mark the subject on the
verb, which is the case for most European languages, a very small number of languages
mark just the object, usually as a result of ergativity or some other complications
pertaining to morphosyntactic alignment, but more than half of all languages with verb
agreement mark both the subject and the object and occasionally the indirect object or other
roles as well, a phenomenon called polypersonal agreement. Verb agreement most often evolves when pronouns
become phonologically dependent on the verb and are then incorporated into it to become
affixes, at which point the verb cannot occur without them. This means that in languages with polypersonal
agreement, most often the verb will be given one set of affixes to mark the subject and
a separate set for the object, and they’ll usually come in the same or a similar order
relative to each other and the verb as they did before they were incorporated, although
once polypersonal agreement evolves, it often allows for increased word order flexibility. Since the pronouns are already expressed on
the verb, their roles can be inferred even if they’re entirely omitted from the sentence. Languages that routinely exclude pronouns
in this way are called pro-drop languages, and although not all pro-drop languages exhibit
verb agreement, the extra information provided by verb agreement makes pronoun-dropping especially
likely. Very often in pro-drop languages, including
the pronoun serves to place focus on it or to highlight its role in the phrase. Somewhat related to pronoun dropping is the
cross-linguistic tendency for a third person singular subject to not be marked overtly,
but instead to be inferred from a lack of marking. Basically, unless anything suggests otherwise,
the default assumption is usually that the subject is third person. The plurality of the verb arguments is most
often inherently expressed with the person markers, since they derive from pronouns that
already reflect grammatical number. But some polypersonal languages also mark
the plurality of either one or both of the arguments elsewhere on the verb as well, a
feature that sometimes gets called “verbal number” or “pluractionality”. In languages like Classical Nahuatl, the verb
takes additional plural marking only for the subject, while in others, like Navajo, the
verbal plural marker can signify either a plural subject or plural object or both depending
on context. Pluractionality most often evolves from frequentative
or distributive aspect marking, signifying that the verb happens multiple times, which
then becomes inextricably associated with the verb having multiple participants, and
so the marking is reanalyzed as a sort of verbal plural. As for the person-markers themselves, In most
polypersonal languages, the subject and object affixes will come in a consistent order, and
which marker refers to the subject and which refers to the object will be interpreted from
their position in the verb template. But occasionally, certain combinations of
subjects and objects will be encoded simultaneously with a single inseparable affix that’s often
not phonologically or etymologically related to the expected subject and object markers. This is called portmanteau agreement, and
while there’s still research being done into how exactly this comes about, there is
some evidence that it evolves when more general person markers take on additional implications
of meaning, such as how in Guaraní, the portmanteau marker that simultaneously encodes a first
person plural subject and a second person singular object seems to have evolved from
an extended use of the intransitive first person plural Inclusive subject marker. Basically “You and I do something” got
reinterpreted as “We do something to you”. And for one final variation on polypersonalism,
let’s talk about direct-inverse systems. In these systems, the positions of the subject
and object markers don’t conform to any pre-determined order in the verb template,
but instead according to what’s called a person hierarchy, which in all attested cases
has either 1st or 2nd person at the top, then 3rd person below those, and then any other
persons the language might distinguish, like inanimate or obviate 3rd person, at the very
bottom. Each of the persons included in the hierarchy
will have an invariable slot in the verb template, regardless of the role they play in the phrase,
which means the identity of the subject and object can’t be inferred from the person
marking alone. Instead, the default assumption is that it’s
the argument higher on the hierarchy doing something to the argument lower on the hierarchy,
and if this isn’t the case, then a special “inverse marker” is added to the verb
to let the listener know the expected roles have been reversed. So how does this sort of system evolve? The answer has to do with salience and animacy. There’s a prevalent cross-linguistic tendency
for languages to place greater focus, either grammatically or syntactically, on animate
arguments rather than inanimate ones, which makes sense considering that humans generally
like to talk about one another, and that inanimate objects don’t really have any agency of
their own or any ability to carry out any independent actions. This is exemplified by languages like Blackfoot,
in which inanimate nouns are simply forbidden from serving as subjects of transitive verbs,
and by languages like Navajo, wherein the nouns in a sentence must come in order of
animacy as dictated by a strict hierarchy, and if this order doesn’t correspond to
the order of the subject and object, then the third person object marker changes to
reflect this. The person hierarchies seen in direct-inverse
languages are an extension of this idea, with animacy distinctions being made among grammatical
persons as well. The higher up on the hierarchy an argument
is, the more “animate” it’s perceived to be, and so the more likely it is to be
the subject. However, if, counter to expectations, the
more animate argument isn’t the one performing the verb, the language may make use of valency
or voicing tricks, like for example the passive voice, to keep the animate argument as the
subject. Over time, as the pronouns then get incorporated
into the verb to become agreement markers, the passive becomes a necessary and mandatory
component of the role-marking system, and so thereby becomes an inverse marker. Those are just some of the ways verb agreement
can manifest, but marking roles solely through verb agreement may open up a few ambiguities
that might need be clarified with other strategies. One common issue that arises is how to interpret
the verb if both arguments are 3rd person. In such a case, how do you tell who’s the
subject and who’s the object? One way to clear this up is to rely on grammatical
gender. If the person-marking on the verb agrees with
the arguments in gender as well, then as long as the arguments have different genders, it
will still be clear who’s doing what to whom. A very common distinction to make is in animate
versus inanimate arguments, As already mentioned, most of the time subjects will be animate
and objects inanimate, and so this distinction may present itself in person-marking. A language like Swahili takes this to a greater
extreme, having around 18 different noun classes, each having a unique verb agreement prefix,
so most of the time the identities of the subject and object are abundantly clear. Alternatively, another strategy for disambiguation
is obviation, which involves two different 3rd person markers, one for the proximate
argument, or the one that’s most central to what’s being discussed, and one for any
other 3rd person arguments, which are termed obviate arguments, which are either less relevant
than the proximate argument or are considered background information. Obviation may be marked on the verb or the
nouns or both. Obviation is very strongly correlated with
direct-inverse systems, as can be seen in the Algonquin languages like Plains Cree and
Ojibwe, which also have an animacy distinction on top of obviation to further minimize ambiguity. Finally, a language with verb agreement may
clarify roles by simultaneously using either of the other two role-marking mechanisms,
word order or noun case. As mentioned earlier, the majority of the
world’s languages have a default word order, and even though having a different role-marking
strategy allows for some extra flexibility, it’s still handy to have a default word
order to disambiguate things if necessary. Classical Nahuatl primarily uses polypersonal
agreement to clarify roles and therefore allows the constituents of a phrase to come in pretty
much any order the speaker wants, but if any ambiguity arises, it falls back on a default
order of Verb-Subject-Object, while Swahili still uses a consistent Subject-Verb-Object
word order despite its pervasive person and gender marking. Combining verb agreement with noun case is
a little bit rarer, but it’s fairly common in European languages like Greek and Latin,
which mark the subject on the verb, but also have a robust case system while keeping word
order pretty flexible. It’s also interesting to note that many
of the Romance languages that descend from Latin have lost the case system outside of
personal pronouns, but have compensated for it with a much stricter word order. Having both noun case and polypersonal agreement
is quite rare, but can be seen in languages like Basque and Georgian, the latter of which
also has a fairly consistent word order as well! All languages are redundant to some extent,
and so may end up double-marking one or more roles to reinforce meaning. For your own conlang, feel free to use a mix
these strategies to employ as much or as little role-marking as makes sense, so long as it’s
always at least reasonably clear who’s doing what to whom. So, in summary, if you want to create a conlang
that features verb agreement: Decide which roles are going to be marked on the verb,
whether it’s the subject, object, indirect object, and/or other roles. Incorporate pronouns into the verb complex
to become agreement morphology. Determine any additional morphology the paradigm
may include, such as pluractionality, portmanteau agreement, or inverse marking. And decide what other role-marking strategies
may be used to eliminate ambiguity. Overall, verb agreement presents the possibility
for a wide variety of unique and interesting grammatical features, and can serve as a very
useful tool for any conlanger’s repertoire.
