Table 1.

Hypotheses for the evolution of complex cues in animal communication, developed in the context of social/sexual plasticity. Evidence for the potential selective advantage of each hypothesis is given

Hypothesis	Theory	Evidence	Possible role in social/sexual plasticity
‘Backup signal’ ‘Redundant signal’	Multiple cues convey one message. The receiver benefits by assessing the message with increased accuracy. The signaller may benefit when the cost of signalling is reduced by spreading investment across multiple components (Møller and Pomiankowski 1993; Johnstone 1996).	Female swordtail fish distinguish hetero- and con-specific males more accurately based on both chemical and visual cues (Hankison and Morris 2003); male wolf spiders use more visual courtship displays when seismic components are inhibited (Gordon and Uetz 2011).	Improved robustness of information transmission in fluctuating social environments (Bro-Jørgensen 2010) and/or accelerated passing of a stimulus threshold (Rouse and Bretman 2016), resulting in phenotypes better suited to the current environment.
‘Multiple messages’	Each cue conveys a different message to one receiver. For example, different sexual ornaments could reflect different aspects of male quality. The signaller and/or the receiver may benefit by increasing the scope of information that can be exchanged (Møller and Pomiankowski 1993).	Components of great tit (Xiphophorus pygmaeus) birdsong are related to different measures of male quality (Rivera-Gutierrez et al. 2010); agonistic male-male signalling in eland antelopes (Tragelaphus oryx) reflect separate aspects of fighting ability (Bro-Jørgensen and Dabelsteen 2008).	Plastic responses can be fine-tuned to multiple features of the environment.
‘Unreliable signal’	Only one cue is a reliable indicator of quality. Any other signals are maintained because they are not costly to produce and are subject to weak Fisherian runaway selection (Fisher 1930). The signaller gains some benefit from the additional, more minor mate preference. The receiver does not gain any increase in the accuracy of the message (Møller and Pomiankowski 1993; Hankison and Morris 2003).	Bill brightness is significantly correlated with male mating success in mallards (Anas platyrhynchos) and plumage only loosely correlated (Omland 1996); female red jungle fowl (Gallus gallus) show a primary preference for male comb colour and weaker preferences for other ornaments (Johnsen and Zuk 1996).	No likely application to social/sexual plasticity.
‘Emergent message’	A single message emerges through the combination of non-redundant cue components. May benefit the receiver by conveying a more general and informative message based on multiple factors (Partan and Marler 2005; Bro-Jørgensen 2010).	Multiple species of songbirds account for a trade-off between trill rate and frequency bandwidth when assessing trills (Ballentine et al. 2004; Illes et al. 2006; Bro-Jørgensen 2010).	Plastic responses can be fine-tuned to multiple features of the environment, and information transmission is more robust.
‘Alerting signal’	One cue component may catch the attention of the receiver and direct it towards one or more other, informative signals. The signaller and the receiver may benefit from improved transmission of the message(s) (Hebets and Papaj 2005; Bro-Jørgensen 2010).	Bornean ranid frog (Staurois guttatus) calls direct the attention of conspecifics towards a visual foot-flagging display (Grafe and Wanger 2007); olfactory signals from male Gasterosteidae sticklebacks may make females alert to subsequent visual signals and increase detection (McLennan 2003).	Informative cues are made more salient, resulting in a phenotype better-matched to the social environment.
‘Receiver psychology’	Complex cues may benefit the signaller and the receiver by enhancing detection, discrimination, learning and memory of the message (Guilford and Dawkins 1993; Candolin 2003; Hebets and Papaj 2005; Bro-Jørgensen 2010).	The presence of auditory signals improves the speed of colour discrimination in domestic chicks (Gallus Gallus domesticus; Rowe 2002); audiovisual stimuli enhance song acquisition and quality in nightingales (Luscinia megarhynchos; Hultsch et al. 1999).	Informative cues are more salient, influential or efficiently processed; as in ‘alerting signal’.
‘Sensory overload’	In agonistic interactions, the signaller may benefit from the transmission of complex cues by reducing the accuracy and/or speed of message transmission (Hebets and Papaj 2005; Bro-Jørgensen 2010).	Dark-eyed juncos (Junco hyemalis) react more slowly when exposed to alarm calls in addition to a visual cue, compared to the visual cue alone (Randolet et al. 2014).	The response of the receiver may be rendered less advantageous or more costly due to time lags (Padilla and Adolph 1996; DeWitt et al. 1998) and phenotype-environment mismatches to the benefit of the signaller.