LINGUIST List 13.899

Mon Apr 1 2002

Sum: Motion Verbs & Manner

Editor for this issue: Marie Klopfenstein <>


  1. Stathis Selimis, Motion verbs + Manner

Message 1: Motion verbs + Manner

Date: Mon, 1 Apr 2002 16:15:32 +0100 (BST)
From: Stathis Selimis <>
Subject: Motion verbs + Manner

I recently posted two questions regarding the coding of motion events 
through verbs. I briefly summarize the most relevant responses below and 
also attach the most relevant comments.

 What criteria should we use in deciding what a motion verb is. Do all 
sorts of movement count, etc.?

It all depends on what investigators look for and why they are interested
 in motion verbs. Some, like D. Slobin, are interested in how people talk
 about moving from one place to another, thus studying verbs of 
translocation. Deciding, however, whether a specific construction has 
translocational meaning or not depends upon contextual factors as is apparent in the 
difference between "he jumped over the fence" -- > translocation, vs. "he
 jumped up and down". This kind of verb polysemy (in Talmy's words (1985)
) can also be treated diffferently: J. Zlatev suggests taking as "potentially 
translocational" all Manner verbs that CAN lead to a change of location 
(both intransitive verbs such as "jump", "bounce", "ride", etc. and transitive 
ones such as "hit", "put", "push", etc. are included). The distinction 
now between intransitive and transitive verbs can reflect the distinction 
between verbs of self-motion and verbs of caused-motion, although these 
two pairs of categories do not always coincide, as K. Jobbagy suggests 
(e.g., "follow" is a transitive verb, however it expresses change of 
location of a moving object, which is the subject of the sentence). Some
 researchers (e.g., Slobin) also include in their interests verbs of conveyance 
("bring", "take", "carry", etc.), because these verbs are at times 
translocational. Actually, these are verbs of a special type, because they 
conflate the self-movement of the "carrier" and the "caused-movement"
 of the thing "carried". On the other hand, Jobbagy seems to limit for her 
particular research purposes translocational verbs as follows in her dissertation 
(in preparation): "[...] Motion verbs are verbs that not only 
implicate but incorporate in their meaning the change of location of a moving 
object, and whose subject is identical with the moving object. From 
this follows that motion verbs must be durative". [Antonopoulou's Ph.D. thesis 
(1987) on Modern Greek Motion Verbs from the angle of Prototype Theory 
includes verbs denoting change of location as well as verbs denoting 
change of posture, both causative and noncausative ones, while Bassea-Bezantakou's 
Ph.D. thesis (1992) on Modern Greek Motion Verbs limits the scope of 
the semantic field in question to those verbs denoting movement of 
the whole Figure (i.e., translocation) which is self-initiated (i.e. the 
Figure is an animate entity).]

 How can Manner be defined, so as we might have clear criteria on whether 
verbs code Manner or Path? In fact researchers treat verbs like "fall"
 as lexicalizing Manner in some studies and Path in other studies.

Manner-of-motion is seen as a multidimentional domain, so one has to decide 
what s/he is interested in. Motor pattern (e.g., "crawl"), rate (e.g.,
 "hurry") and attitude (e.g., "stroll") are considered to be indicative of
 Manner (Slobin). As J. Zlatev reminds us, Manner corresponds to one of 
Talmy's (2000) Co-events, which accompanies a main event. This can be shown 
by analyzing a compound event such as "He glided down" into two sub-events 
"(a) He moved downward (b) in a gliding manner", the second one denoting 
how the movement (rather, change of location) is performed. However,
 Manner, in the broad sense, can also encompass Speed, Vehicle, Means, etc. 
(Zlatev). [At this point, I remind of Antonopoulou's dissertation (1987), 
where Manner 'is used in a fairly broad sense and is meant to cover three 
different areas: a central one involving various types of (typically) 
human motion on ground, e.g. "trexo" (run), "perpatao" (walk), a much wider 
area involving different ways of moving in water or air, e.g. "kolibao" 
(swim), "petao" (fly), and a most restricted one involving distinctions 
at a low level of inclusiveness as exemplified by hyponyms of verbs belonging 
to the central area, e.g. "vimatizo" (pace), "dhraskelizo" (stride)' 
(pp. 237-238). Antonopoulou also writes (p. 242) that '"medium", 
"instrumentality" and "impetus" are regarded as being most closely related to
 "manner" in the sense that their interaction results in various types of
 locomotion", and she adds that a lot of "Manner" distinctions depend on 
"type of object" moving (e.g., "stazo" (drip)), "speed" (e.g., "arghokilao" 
(flow slowly)), "length of distance covered" (e.g., "porevome" (walk a
 long distance)), and combinations of "speed and length" (e.g., "vradhiporo" 
(walk a long distance slowly)). Bassea-Bezantakou (1992), in defining
 Manner, takes into account factors such as type of movement perceived, 
instrument (e.g., fly), spatial (e.g., what the ground consists of) and 
temporal (e.g., speed) specifications, (marginally) intentionality, and the
 psychological, physical and social conditions of the Agent.]

As far as the specific question on "fall" is concerned, it tends to be 
treated as a Path verb (as it simply means to move downward without 
self-control according to Slobin, among others), but even Slobin notes that 
considering it as a possible Manner verb does not seem so far-fetched. 
According to H. Harley as well, "fall" encodes just Path, while "Manner 
crucially involves some element of meaning that remains constant throughout the 
event; an element that is not, in Vendlerian terms, evolving towards a 
culmination point, but is just 'accompanying' the event over its duration. 
The problem with 'fall' is that for sufficiently unbounded falling (e.g.,
 the weightlessness of astronauts in orbit), there is no culmination 
point, at which point it might be a Manner verb". Zlatev on the other hand 
solves the problem by considering 'fall' (and 'sink' as well) an 
in-between case which conflates an element of Manner (in the broad sense) with
 Path. He believes that such verbs place constraints on Path and thus 
encode Manner in this general sense.


Note: I am grateful to the following contributors for their inspiring ideas.

Alexis Dimitriadis,

Anetta Kopecka,

Dan I. Slobin, slobinsocrates.Berkeley.EDU

Daniel Loehr,

Andrea Schalley,

Heidi Harley,

Jess Tauber,

Jordan Zlatev,

Katalin JobbE1gy,

Paul Chilton,

Robinson Ed, & Nancy Felson,

Teenie Matlock,

William L. Jarrold,

St. Selimis

Main points from some of the responses received:

From: Jess Tauber, His approach is based on ideophones and
 expressives: "These were not covered by Talmy in his own work, yet they 
are likely of singular importance in any historical work dealing with manner vs. path 
evolution. Some languages have many, many thousands of
these items- Japanese, Korean, many Mon-Khmer, almost all Bantu and 
related Niger-Congo, etc. Others show definite traces of
relatively recent mass lexicalization of these, leaving a much reduced set 
of free forms- Mongolian, Turkic, Uralic, Tungusic. Later
 even fewer free forms exist, but many more lexicalized ones- European IE, 
Austronesian, Tai-Kadai, etc. There seems to be a general
 trend towards higher ranking insertion priveleges as one proceeds along 
this continuum- both within the lexicon proper and higher
up- the more grammatical affixation, the fewer free ideophones. Eskimo, N
WCaucasian, and other extensively polysynthetic
languages have almost none at all. As mentioned earlier, such items gener
ally encode manner of activity. Generally free ideophones
 have a semi-punctual type of aktionsart, so pathway is never extensive. 
Only when reduplicated does path semantics become more
 obvious, and that is by addition (for example, arcs to multiple circles)
. Ideophones also tend to be organized paradigmatically, in that
similar forms encode similar notions in regular and predictable ways. 
Interestingly, pure pathway terms tend to have an opposite
developmental implicational hierarchy, so that their greatest expansion (
into large-scale paradigms) is in just those polysynthetic
languages that have the fewest ideophones. Anyway, most languages are in 
between in terms of their lexical and syntactic organization.
Some show, like IE, both manner and pathway elements fused to the verb st
em- but the focus can clearly be shifted, as is the case
between Romance and Germanic, for instance. Interestingly here as well, 
one can actually see the proportionality of onomatopoeic
free or bound, lexicalized forms shift in synch. [...]"

From: Teenie Matlock

Suggested reading: Miller, G.A. (1972). English verbs of motion: A case 
study in semantics and lexical memory. In A.W. Melton & E. Martin (Eds.), 
Coding processes in human memory (pp. 335-372). New York, NY: John Wiley 
& Sons.
Miller, G. A., & Johnson-Laird, P.N. (1976). Language and perception. Cam
bridge, MA: Harvard University Press.

From: Anetta Kopecka

It seems to me that in Talmy`s work there is not ambuguity that "put" 
expresses motion. I think that he considers "put" as a positional motion 
verb that conflates "cause" component but not manner. If I understood his 
interpretation "put" encodes a translational motion excerced and controlled
 by an agent untill the end-point of the event. It`s a goal-oriented 
motion verb. "take" and "bring" are more deictic. [...] It seems to me that 
manner can be paraphrased, for example: I run to school 3D I went to 
school running.

From: William L. Jarrold,

In the Cyc ontology, they have encoded all of the sorts of distinctions 
you are talking about. [...] We have an overall broad concept called 
#$MovementEvent which would include rotation, translation and movement of a 
part (e.g. a facial movement), causing other objects to move (e.g. putting). 
And then there are distinctions for each of the above specializations 
of movment (e.g. there are separate conceptual nodes for rotation, 
translation, movement of a part, etc.) and there are many more finer gradations.
 There are natural language hookups to these conceptual nodes. You might 
look up Beth Levine's book on verb classes. She probably has something to
 say on the matter.

From: Dan I. Slobin, slobinsocrates.Berkeley.EDU

There is no "right" answer to your questions. It all depends what you're 
looking for, and why you're interested in motion verbs. In my work I'm 
interested in how people talk about moving from one place to another, so I 
study verbs of translocation, and in constructions that have translocational 
meaning (e.g., "he jumped over the fence" but not "he jumped up and down"; 
"he ran to school" but not "he ran around in the yard"). In many 
languages these two types are distinguished in the grammar--as in the choice 
of 'be' (translocational) vs. 'have' (activity) as auxiliary. I also 
distinguish between verbs of self-motiion (intransitive) and verbs of cause
d-motion (transitive)--because these have different argument structures, 
and because languages often have different set of verbs for the two sorts. 
I include verbs of conveyance ('bring', 'take', 'carry') in my analyses, 
because these verbs are translocational; but I also treat them as a special 
type, because they conflate the self-movement of the 'carrier' and the 
caused-movement of the thing'carried'.
As for manner-of-motion, this is a multidimensional domain, and, again, 
you have to decide what you're interested in. In my work, I attend to the 
full range of verbs that encode motor pattern (e.g., 'crawl', 'walk', 'run'), 
rate (e.g., 'hurry', 'dash'), and attitude (e.g., 'stroll', 'amble',
 'saunter'). I debated a long time about 'fall', but finally decided that
 it is a pure change-of-location verb, because there is no particular mot
or pattern, rate, or attitude involved. 'Fall' simply means to move downward 
without self-control. But I can see why one might also want to consider 
'fall' a manner-of-motion verb."

From: Jordan Zlatev,

I agree with you that there is much less constisteny than we would wish in 
semantic treatments of Motion. One of the influential studies, Talmy 
(1985)'s "Lexicalization patterns", left both questions you raise essentially 
unresolved. But I am pleased to see that the revised version in Talmy 
(2000) is rather clearer. In relation to your questions, the position seems 
to be:1) A Motion event (capital M) is a "situation containing motion 
and the continuation of a stationary location" (25), but this VERY general 
category can be devided into "translational" ("an object's basic location 
shifts from one point to another in space" and "self-contained motion"
 - "an object keeps its same basic, or "average" location" (35). The problem 
with deciding about particular verbs is that Motion (in both the general 
and translational sense) and what is now called the "Co-Event" (that 
can involve Cause, Manner, Enablement etc) are regularly "conflated". So 
your examples "bring", "take" and "put" express Motion + Cause (possibly 
also Deixis for the first two). The problem is that many "manner verbs" 
like "float" and "kick" can appear in both a translational motion and in a
 locational context: "Hi kicked the ball/the wall". Talmy tries to explain 
this with polysemy - there are 2 different senses, where the first is 
translational, the second locational. But this is clearly problematic, and
 even Talmy adds that a constructional approach a la Goldberg (1995) may 
be prefered. My approach is to treat as "potentially translocational" all
 manner verbs that CAN lead to a change in location: intransitive like 
"jump", "bounce", "ride", "roll" and transitive like "hit", "put", "blow", 
"push". In working with a corpus, the decision on whether translational 
motion is involved will have to be worked out from the context. In the case 
of the "path verbs" in Romance etc. the situation is clearer since they
 are per defnition translocational (I prefer this term than "translational").
2) "Manner" too can be defined more generally to mean something like 
Talmy's "Co-event" - and then can be subcategoried with respect to the 
various "relations" that the "co-event" stands to with respect to the (main) 
translocational event, where "manner" in a narrow sense is but one: "He
 glided down" 3D "He moved downward in a gliding manner"Other types of 
"Manner" in the broad sense involve Spead, Vehicle, Means etc. My main 
reservation with respect to Talmy is the apriorisic decision to call 
translocation, or more generally "the core schema" for the "main/framing event" 
and all other type of information "the co-event". I can not see the basis
 for such a generalization accord the board, but only in say, the Romance
 languages. As far as verbs like "fall", "sink" etc. verbs are concerned,
 I think they take an intermediary position between the "pure" manner 
verbs (that could be locational in some contects at least - "He walked for 
hours on the treadmill") and the path verbs like "enter" and "exit". I 
think they conflate an element of Manner (in the broad sense) and of Path 
too. My criterion is that they place constraints on Path, so that e.g. 
*"I fell upwards" is wierd. So in my work I code them seperately, calling 
them Manner+Path (MP) verbs. Interestingly they also take a seperate 
structural position in the language I work with, Thai.

From: Daniel Loehr,

Check out: (WordNet)

From: Andrea Schalley,

1) I guess this depends on the reference you assume. That is, with regard
 to the body, the hand is moving while putting things. With
 regard to the background in front of which the body as a trajector exists, 
it probably doesn't (or maybe the hand does, but the slight
 facial movement doesn't). So it seems to be a fuzzy distinction. Nevertheless, 
my intuition tells me that for a verb to be a motion verb
there has to be a movement of the whole body involved. This movement can 
be along a path (run) or at a location (jump). So I
wouldn't include `to put something' into the class of motion verbs as long 
as only a movement of an arm or a hand is necessary. [...] Here I agree 
for bring and for the reading of take which includes a movement to the 
place of reference (i.e., the place the thing or
person has to be taken to, as in taking something to the kitchen (while in 
the living room)). If I stand in front of a table and take a
book from the table, I would not consider take to be a motion verb. 2). 
[...] concerning your example [fall] I would prefer the Manner reading 
(just intuitively - maybe because a Path analysis is not
distinctive enough, as flow, run, fall, jump, etc. would all entail Path 

From: Heidi Harley,

1). I have always understood motion verbs to be those verbs that involve 
motion or change of position of the body as a whole. "sit" and "stand", in 
the active senses, then, would be motion verbs, "sit" and "stand" in the 
stative sense ("The lamp stands in the corner"), verbs describing facial 
movements and placing things with the hands (without moving the whole body) 
would not be. "Bring" and "take", insofar as they imply movement of 
the bearer along with the Theme, are motion verbs. (Pinker has some 
interesting remarks in his 1989 book on the availability of the double-object 
construction in English with such verbs -- although I think 'bring' and 
'take' are counterexamples to his claim, which is that if the bearer has to 
go along with the Theme, it's not dative-shiftable: "I threw Mary the 
ball" but "??I schlepped Mary the box".

2). I would think "fall" encodes Path, not manner. Manner, I think, crucially 
involves some element of meaning that remains constant throughout the 
event; an element that is not, in Vendlerian terms, evolving towards a 
culmination point, but is just 'accompanying' the event over its duration
. The problem with 'fall', I suppose, is that for sufficiently unbounded 
falling (e.g. the weightlessness of astronauts in orbit), there is no 
culmination point, at which point I suppose it might be a manner verb.

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue