FOR PROF. GANNDHI -- please don't spam !!

science vocabulary.’ American Educational Research Journal 24, 541–545.

Stahl S, Duffy-Hester A & Stahl K (1998). ‘Everything you wanted to know about phonics (but were afraid to ask).’ Reading Research Quarterly 33(3), 338–355.

Tinkham T (1997). ‘The effects of semantic and thematic clustering on the learning of second language vocabu- lary.’ Second Language Research 13, 138–163.

Waring R (1997). ‘The negative effects of learning words in semantic sets: a replication.’ System 25, 261–274.

Waring R & Takaki M (2003). ‘At what rate do learners learn and retain new vocabulary from reading a graded reader?’ Reading in a Foreign Language 15, 130–163.

454 Vocabulary: Second Language

Vocal Production in Birds

G J L Beckers, Leiden University, Leiden, The Netherlands and Max Planck Research Insitute for

Ornithology, Seewiesen, Germany

� 2006 Elsevier Ltd. All rights reserved.

Bird vocalizations are widely noted for their extreme diversity in acoustic structure and sometimes high degree of complexity. Understanding the mechanistic basis of vocal performance in birds has been a long- standing effort. This is largely because of intrinsic scientific interest, but also, more recently, due to the growing importance of birdsong as an animal model system for the development and neurobiology of human speech. In this article, I review some highlights of current knowledge on the peripheral mechanisms underlying vocal production in birds and point out notable parallels and differences with speech production in humans.

Principles

The biophysical principles of vocal production in birds are similar to those involved in voiced speech of humans. Air flow, induced by the action of respira- tory muscles, is modulated by vibrating structures in a voice organ, a process that causes acoustic pressure waves. These waves propagate through the air spaces of the vocal tract and eventually radiate into the outside environment. The exact characteristics of the sound that is produced depend on the vibration properties of vocal source, the acoustic resonances of the vocal tract, and the acoustic radiance function of the structure from which the sound eventually is emitted, e.g., the beak or ‘vocal sacs.’ To a first ap- proximation, vocal production in birds can be explained, like human speech, by considering the properties of the vocal source and vocal tract inde- pendently, even though they function as one system and interactions between these two components may sometimes occur.

Vocal Source

Voice Organ

Vocal source vibrations in birds are generated by a unique voice organ, the ‘syrinx,’ which is an evolu- tionary innovation not found in any other group of animal. Birds also possess a larynx, which is the voice organ in mammals, reptiles, and anurans, but the larynx has not been found to be involved as a sound source in bird vocalization. The sole function of the syrinx is sound generation, and from a vocal produc- tion point of view it may therefore be less constrained in its mechanical design than the larynx, which also has functions related to respiration and feeding. For this reason perhaps, many birds appear to be much better vocal performers – at least from a pro- duction point of view – than any laryngeal vocalizer, including humans.

The syrinx consists of a modification of the airway at the junction of the trachea and bronchi and is situated in an air sac deep inside the thoracic cavity. In general, it comprises a cartilaginous or bony framework containing vibratory structures, either membranes or labia, and muscles that control the geometric configuration of the framework. The spe- cific morphology of the syrinx, however, may differ considerably between bird taxa, even in basic design.

The most complex syrinx is found in the most virtuous singers among birds: the songbirds (suborder Oscines or Passeres, comprising approximately 40% of all bird species). The songbird syrinx is special because it has two sound sources, a voice at the cranial end of each bronchus, which can be controlled independently by highly specialized musculature (Figure 1).

The bipartite nature songbird of the syrinx is exploited in various ways across taxa to enhance vocal performance in spectacular ways. For example, in many songbirds, each syringeal voice specializes in its own vocal register – the left voice usually covers a

Figure 1 The songbird voice organ. This bipartite syrinx is situated at the junction of the bronchi and trachea. Schematic

horizontal section illustrating the two voice structures, each con-

sisting ofa lateral labium(LL)andamedian labium(ML)at the top

ends of the bronchi. BC, bronchial cartilage; 3B, third bronchial

cartilage;BL,bronchial lumen;dS,m.syringealisdorsalis;dTB,m.

tracheobronchialis dorsalis; ICM, membrane of the interclavicular

air sac; M, syringeal muscle; MTM, median tympaniform mem-

brane; SM, semilunar membrane; ST, m. sternotrachealis; T, tra-

chea; TL, m. tracheolateralis; TY, tympanum; vS, m. syringealis

ventralis; vTB, m. tracheobronchialis ventralis. � Illustrations by Claus Bo Næsbye Larsen. Source: Larsen O N & Goller F (2002)

Journal of Experimental Biology 205,25–35.TheCompanyofBiologists

Ltd. Reproduced with permission.

Vocal Production in Birds 455

lower range than the right voice. Songbirds such as the Northern cardinal, Cardinalis cardinalis, use this system to produce seemingly continuous notes modulated over extreme frequency ranges, by starting phonation on one side of the syrinx and quickly switching to the other side in the middle of the note. Other birds such as the brown thrasher, Toxostoma, rufum, use the two-voice system to simultaneously produce unrelated sounds and thus sing a duet in solo. A totally different approach to achieve vocal virtuosity using a bipartite syrinx is taken by Water- slager canaries, Serinus canaria, which can sustain high syllable-repetition rates, sometimes exceeding 30/s, for long periods of time, sometimes exceed- ing 30s, without taking a pause for a normal inspira- tion. This is achieved by using the left syrinx side for phonation while the right side is used to take so- called mini-breaths in between each syllable. Such a system allows the left side to be kept in phonatory position while the right side is able to maintain respiratory air supply for phonation, a specialisation that increases the duration of the period in which fast syllable repetition can be sustained.

Less complex or even relatively simple syrinxes are found in nonsongbird taxa, which, overall, comprise less virtuous vocalizers. However, attempts to find a clear and systematic relationship between syringeal

complexity and phonetic complexity in birds have failed thus far.

Voice Mechanism

The actual mechanism by which sound is generated in the syrinx has been a long-debated topic. Normally, voice production in tetrapods involves a pneumatic valve-type sound generator, such as the laryngeal voice in mammals, in which vibrating vocal folds periodically close and open the airway lumen, causing a train of air pulse releases in the vocal tract. Such a system characteristically produces a sound with a fundamental frequency and a series of harmonic over- tones. However, birds often produce pure-tone song, which lacks the characteristic stack of harmonic over- tones of a valve-type source. For this reason, a variety of alternatives to a valve-type sound generator have been proposed for pure-tone bird vocalization, the most notable of which is an aerodynamic whistle mechanism. However, recently it has been shown that pure-tone song in doves (nonsongbirds) is gener- ated by the syrinx as a sound with harmonic over- tones, which is subsequently filtered by the vocal tract to a pure tone before the sound is emitted. Moreover, direct video observation of the syrinx in phonating birds shows that vibrating membranes or labia form a slot in the airway lumen. Pure-tone phonation, as well as other types of phonation observed in birds, can therefore in principle be explained on the basis of a normal valve-like source mechanism in the syrinx, similar to laryngeal voice mechanisms in humans. Nevertheless, given the extreme diversity in syringeal structure across species, it remains possible that fun- damentally different mechanisms of syringeal sound generation exist in birds.

Modulation of Voice

Source phonation in birds is usually modulated in time to generate complex spectrotemporal vocal pat- terns. There are two fundamentally different ways in which this is achieved.

First, coordinated neuromuscular control of the respiratory and syringeal systems can directly modulate biophysical characteristics of the voice system, so that spectral patterns and overall amplitude of the sound that is produced at the syrinx change over time. The fundamental frequency of phonation in songbirds, for example, changes by the action of ventral syringeal muscles, which presumably regulate the tension of the vibrating labia, whereas in dove coo phonation, fre- quency is modulated by changes in air sac pressure achieved by respiratory muscles. Likewise, vocal ampli- tude is known to be controlled by a variety of direct neuromuscular mechanisms in birds.

456 Vocal Production in Birds

A second and relatively little known mechanism of modulation arises spontaneously from intrinsic properties of the vocal production system. The pri- mary sound generators in birds, vibrating labia or membranes, are in fact nonlinear oscillators, and it has been observed in diverse scientific fields that coupled nonlinear oscillators in general can also exhibit complex dynamics without any external, complex control (nonlinear dynamics or ‘chaos the- ory’). The key point about nonlinear dynamics in vocalization is that a relatively simple system, such as two coupled vibrating membranes, can give rise to a series of different states of oscillatory behavior, which include amplitude and frequency modula- tions. Most likely, at least the extremely rapid peri- odic modulations found in bird vocalizations, which sometimes exceed 500 Hz and are unlikely to be under direct muscle control, originate from such intrinsic dynamics, as do discontinuous mod- ulations such as the often observed frequency jumps (‘voice breaking’).

Most bird vocalizations that are described as ‘noisy’ are likely also due to nonlinear dynamics and may be produced by a chaotic oscillatory state of the syrinx. It is interesting to note that this is in contrast with noisy sound sources in human speech, such as whispered consonants (e.g., English h), which do not originate from vocal fold oscillations but from air turbulence in the vocal tract, a process that indeed is truly random or noisy. Chaos, however, is completely deterministic and only sup- erficially resembles noise in spectrographic repre- sentation. Nevertheless, some bird vocalizations may actually originate from a true noise source, at least as judged by the human ear, for example, hissing in swans.

Vocal Tract

The sound that is generated by the syrinx depends not only on its oscillatory dynamics, but also on the specific acoustic impedance at the syrinx. That is, source vibrations at some frequencies produce sound more easily than those at other frequencies, a pattern that is determined by the geometry of surrounding vocal tract air cavities such as the bronchi and trachea. Moreover, as sound propagates through the vocal tract it may undergo further filtering through changes in tract diameter, which cause acoustic im- pedance changes. Thus, the avian sound production system should be viewed as a source–filter system, in which the syrinx provides the necessary source vibra- tions and the filter determines the amplitudes of the sound frequencies produced.

Although acoustic filtering of syringeal source vibrations is physically inevitable, it has been a matter of long-term debate how important vocal tract filter- ing actually is in shaping specific time–frequency pat- terns in bird vocalization and which mechanisms are used to shape vocalizations. Three models of vocal tract filtering have been suggested in the literature, each of which is likely to occur in birds.

In the simplest model, the vocal tract could act as a static broadband filter that leaves the relative ampli- tudes of all frequencies produced by the syrinx more or less unaffected. In an alternate version of this model, the vocal tract could act as a static bandpass filter with a bandwidth that is smaller than the fre- quency range in which the syrinx produces sound (Figure 2, ‘Static filter’). The latter mechanism approximates the situation in dove coos, where har- monic overtones produced by the syrinx are filtered out so that only the fundamental frequency remains. Such a mechanism can be used only when relevant sound features, in the case of doves the fundamental frequency, are modulated within the existing static vocal tract pass-band.

In a second model, the frequency range over which the syrinx modulates relevant sound features is much larger than the bandwidth of the vocal-tract filter, in which case the vocal tract must dynamically adjust resonance frequencies in order to track the changes in phonation frequency (see Figure 2 ‘Dynamic filter tracking source’). This mechanism is more complex because the dynamic tracking requires close motor coordination between the syringeal source and vocal tract articulators. However, the vocal tract does not create complex phonetic patterns independently of the source. As an example of this mechanism, it has been suggested that the beak gape changes that often occur during bird vocalizations adjust the frequencies of vocal tract resonance peaks so that they follow a particular harmonic frequency produced by the syr- inx. Recent research has shown that beak gape changes do indeed modulate vocal tract filtering char- acteristics, but not through the shifting of resonance peaks. This notwithstanding, the argument may hold for other vocal tract articulators.

Finally, analogous to human speech production, a dynamic vocal tract filter can create its own complex time–frequency patterns, independently of source modulations generated by the syrinx (Figure 2, ‘Dynamic filter with independent modulation’). This makes sense only if the syringeal source produces a wide-band sound, such as a harmonic stack. An example of such a mechanism occurs in parrots, whose tongue movements cause significant and com- plex changes in four vocal tract resonances or ‘for- mants.’ Analysis of parrot mimic speech and natural

Figure 2 Three models of vocal tract filtering in bird vocaliza- tion. Subfigures are schematic spectrographic representations.

In the toprow, the frequencyrangesofvocal tract resonancesare

indicated as gray shading. The harmonic sound frequencies pro-

ducedby thesyrinxare indicatedas thick lines,continuous if they

overlap with vocal tract resonance ranges and dashed if they fall

outside these ranges. The bottom row of subfigures shows the

sound that is thus emitted from the bird as vocalization for each

model. It should be noted that the vocal tract resonance patterns

shown are simplified. In reality, the avian vocal tract normally

has more than one or two resonance peaks.

Vocal Production in Birds 457

vocalizations suggests that parrots indeed use tongue placement to modulate formant patterns indepen- dently of syringeal source modulations, which is an interesting extension of the already known parallels between human speech production and bird vocal production.

In conclusion, there are notable differences be- tween vocal production in birds and that in humans, the most important of which relate to the fact that birds have evolved an innovative voice organ that is, at least in songbirds, extremely versatile. Neverthe- less, the parallels in basic sound generation mechan- isms and a source–filter system are equally notable. This is especially true for species that use vocal tract filtering to create complex modulation patterns in- dependently of the voice source, a fundamental mechanism in coding information in human speech.

See also: Animal Communication Networks; Animal Com-

munication: Overview; Animal Communication: Vocal

Learning; Birdsong; Communication in Grey Parrots; De-

velopment of Communication in Animals; Dialects in

Birdsongs; Individual Recognition in Animal Species;

Production of Vocalizations in Mammals.

Bibliography

Beckers G J L, Nelson B S & Suthers R A (2004). ‘Vocal- tract filtering by lingual articulation in a parrot.’ Current Biology 14, 1592–1597.

Beckers G J L, Suthers R A & ten Cate C (2003). ‘Pure-tone birdsong by resonance filtering of harmonic overtones.’ Proceedings of the National Academy of Sciences USA 100, 7372–7376.

Fee M S (2002). ‘Measurement of the linear and nonlinear mechanical properties of the oscine syrinx: Implications for function.’ Journal of Comparative Physiology A 188, 829–839.

Fletcher N H & Tarnopolsky A (1999). ‘Acoustics of the avian vocal tract.’ Journal of the Acoustical Society of America 105, 35–49.

Gaunt A S & Nowicki S (1998). ‘Sound production in birds: Acoustics and physiology revisited.’ In Hopp S L, Owren M J & Evans C S (eds.) Animal acoustic communication. Berlin: Springer. 291–321.

Goller F & Cooper B G (2004). ‘Peripheral motor dynamics of song production in the zebra finch.’ Annals of the New York Academy of Sciences 1016, 130–152.

Goller F & Larsen O N (2002). ‘New perspectives on mechanisms of sound generation in songbirds.’ Journal of Comparative Physiology Series A 188, 841–850.

Greenewalt C H (1968). Bird song: Acoustics and physiol- ogy. Washington, DC: Smithsonian Institution Press.

Hoese W J, Podos J, Boetticher N C & Nowicki S (2000). ‘Vocal tract function in birdsong production: Experimen- tal manipulation of beak movements.’ Journal of Experi- mental Biology 203, 1845–1855.

King A S (1989). ‘Functional anatomy of the syrinx.’ In King A S & McLelland J (eds.) Form and function in birds. London: Academic Press. 105–192.

Laje R, Gardner T J & Mindlin G B (2002). ‘Neuromuscu- lar control of vocalizations in birdsong: A model.’ Physi- cal Review E 65, 051921.

Larsen O N & Goller F (2002). ‘Direct observation of syringeal muscle function in songbirds and a parrot.’ Journal of Experimental Biology 205, 25–35.

Nelson B S, Beckers G J L & Suthers R A (2005). ‘Vocal tract filtering and sound radiation in a songbird.’ Journal of Experimental Biology, 208, 297–308.

Suthers R A (1999). ‘The motor basis of vocal performance in songbirds.’ In Hauser M D & Konishi M (eds.) The design of animal communication. Cambridge: MIT Press. 37–62.

Suthers R A & Zollinger S (2004). ‘Producing song: The vocal apparatus.’ Annals of the New York Academy of Sciences 1016, 109–129.