beats with about three or four strong beats per unit,and all under
an intonation contour (melodic contour) where the end of the unit is
signaled by tonal changes.Finally,both human conversational formulas
and gelada vocal units are produced with no hesitation phenomena and
no internal pauses as one continuous gushing-forth of a whole unit of
activity.
Both species use rhythm and melody to distinguish different kinds of
utterance types.Geladas,for example,appear to signal the main differ-
ence between quick back-and-forth alternations and the long friendly
series by rhythmic and melodic features.The long friendly series is always
ended by sustained,stepwise rising tones that seem to indicate the
completion of a series (see Richman 1976,1987).In human speech,
intonational features signal differences among questions,assertions,and
commands.
The gelada productive unit for friendly vocalizing is quite variable in
its internal details of length,tempo,rhythm,musical intervals,and sylla-
ble types,even though the overall rhythmic and intonational contours
are generally similar for all units (Richman 1976).Such variability from
unit to unit again makes these vocal units similar to human vocal
formulas.
The source of the variability of internal details of units for geladas
seems to be the presence of layer upon layer of different expressive
features by means of which the animals express overlapping,conflict-
filled motivations and emotions they bring to each friendly encounter
(Richman 1996).These expressive features are carried by vocal articula-
tory gestures that are quite distinctive acoustically as well as visually.
Such overlapping features lead to the presence in gelada units of a great
variety of syllables with different consonant and vowellike features,as
well as a great variety of rhythmic and melodic variation.
In one major respect,however,gelada vocal units offer a striking con-
trast to human vocal formulas:they are not formulaic at all;that is,a
given sequence does not tend to be repeated as a vocal formula.In other
words,we have no evidence that geladas are capable of repeating the
same,exact succession of vocal features that would mark two vocal units
as the same.At most,they can repeat or vocally match specific phrases,
up to three syllables long,that they have just heard.This is in stark con-
trast to people,who in normal,everyday conversation frequently and
effortlessly repeat exactly entire long formulas they have just heard
(Tannen 1989).This contrast,I suggest,provides a significant clue to the
nature and origin of human language.
The particular patterns of syllable types,rhythms,and melodies that
we find in gelada sequences seem to reflect temporary,real-time chang-
ing motivations and interactions of the participants.No special formu-302 Bruce Richman