priately called lexical syntax or lexicoding.The lower level,with mean-
ingless sounds combined into sequences,may be termed phonological
syntax or phonocoding.Meaningful sentences require lexical syntax.
Phonological syntax concerns the rules for sequencing,not the genera-
tion of meanings.Does evidence exist that either of these steps toward
language has been taken by animals?
We can begin with the sentence and work down in reductionistic
fashion.Some animal sounds do posses symbolic meanings,but although
in some cases animals string symbolically meaningful calls together in
the course of their natural vocal behavior,I know of no case in which a
string qualifies as a sentence.Aside from marginal cases (e.g.,Mitani and
Marler 1988),we do not seem to have any recorded natural example of
an animal unambiguously satisfying the crucial criteria for lexical syntax.
No naturally communicating animal is known to sequence symbolically
meaningful calls to make a sentence that has a new,emergent meaning
derived from the combined meanings of its assembled parts.
So much for lexical syntax.How about words and phonemes,or their
equivalent? Symbolically meaningful animal signals such as alarm and
food calls of monkeys and birds all seem to come as indivisible packages.
It is true that their meanings are not completely fixed and immutable,
and can be modulated by giving calls singly or repeatedly,quickly or
slowly,loudly or softly (Marler 1992).But their basic indivisibility sug-
gests no obvious analogue to phonological syntax in their construction.
However,if we widen the search to embrace not only animal vocaliza-
tions with symbolic meanings,but also those of a more classical,affec-
tive kind,impoverished in referential content,but rich in emotional
content,we find something very different.Here are many cases of phono-
logical syntax (Ujhelyi 1996).In particular,scrutiny of the literature on
the structure of learned birdsongs reveals case after case of birds that
employ phonocoding to create individual song repertoires numbered in
the hundreds.These repertoires are generated by reusing again and
again,in many different sequences,a basic set of minimal acoustic units,
the bird’s equivalent of phonemes and syllables.I will limit myself to two
examples,one with a very small individual song repertoire,the swamp
sparrow,the other,the winter wren,with a larger one.
The common song of the swamp sparrow is a simple case.Each male
has two or three songs,each consisting of a two-second string of repeated
syllables,uttered in the spring and summer,many times each day (Marler
and Pickert 1984).The natural songs of this species have many syllable
types.Each syllable,repeated in identical fashion to form the song,is
made up of two to six different notes in many different combinations.
The notes themselves are meaningless,but assembled into distinctive
clusters,they form the basic building blocks of swamp sparrow song.37 Origins of Music and SpeechMUS3 9/14/99 11:58 AM Page 37