Communication involving ritualized displays or graded signals is used to confirm or reject information received from others. Contexts in which this occurs varies considerably: general social interactions, indications for species, sex, and sexual receptivity, situations pertaining to relative social status, negotiation of social interactions et cetera. Communication can be instantaneous as is the case with tactile behaviors and visual displays. These behaviors are also relatively "short-distance" signals. Vocal communication is instantaneous, but may reach over longer distances. Olfactory and pheromonal signals provide information that can be assessed over distances and across time. Dogs use all of these modalities, and we can best understand what the dog is telling us when we have sets of information that are congruent (e.g., the vocal signal matches the physical behavior). Extremely important signals are always redundant, so that more than one modality is involved in the signal. Such patterns act over evolutionary time to minimize costly errors.

To understand canine vocal behavior behaviors it is critical to understand the component signals as they are used to communicate with conspecifics (see Table).

Canine Signals.

Vocalizing/barking is an excellent example of a canine behavior that is complex and best interpreted within the context of the circumstances and the dogs other behaviors. Human response to the barking becomes part of the context.

Dogs do not continuously bark in response to every stimulus, nor do they use the same bark in all circumstances. We can measure aspects of vocalizations by using sonographs - voice spectrographs - of the vocalizations. Analysis of dog barks has revealed that, regardless of their experience with dogs, the vast majority of humans would correctly identify a dog who was vocalizing in a fight and one who was vocalizing when a stranger approached.¹ Less agreement was found for dogs who vocalized on walks, in play, when alone or for those given a ball. While people may be confident that they recognize a frank aggressive or alert response, we should be aware that they may not know that they are not accurately identifying the information contained in barks in other situations. This is important because the information contained in vocalizations across such situations is different.

One study examined growls in 3 contexts (play, when guarding a bone and when approached by a threatening stranger).² Play growls differed from growls in the two agonistic contexts with respect to frequency and formant (spectral/amplitude peaks in frequencies) dispersal: play growls had higher frequencies and lower formant dispersal. While lower frequency growls usually correlate with larger dogs, it's important to note that growl frequency has a functional component which can be altered by contracting the muscles of the throat. Formants can also be functionally altered: when a dog stretches his neck, he behaviorally lengthens his vocal tract and disperses formants, making him appear 'larger' to someone who has made the general association between these vocal characteristics and dog size. This information content is recognized by and affects dogs.³ When dogs' heart rates are measured in response to a series of auditory cues, they significantly increase after hearing a recording of a dog barking at a stranger.

Classes of barks and other vocalizations are consistent across different dogs, given the specific context.⁴ When 10 pet dogs were tested in 3 contexts (play, isolation and disturbance) the dogs vocalized in ways that were similar, showing that the signature of the bark was associated with the context. This pattern indicates that dogs are communicating useful, context-dependent information. We should learn from such information.

Similar patterns for vocal signatures were found when barks from dogs with separation anxiety were compared with barks from unaffected dogs.⁵ The barks used by unaffected dogs in response to outdoor stimuli indicated an alert, similar to the high frequency, solitary note shown by dogs in play, but the repetitive, atonal, complex barks exhibited by dogs with separation anxiety resembled the vocalizations in both the disturbance and isolation contexts. Dogs can clearly recognize the information provided in such signals, and for us to understand dog behavior, we must do so, also.

Humans have been shown to classify the 'emotionality' of canine vocalization with respect to tone, pitch and inter-bark interval.⁶ If both formant dispersions and fundamental frequencies are manipulated, lower formants and fundamental frequencies are rated as belonging to larger dogs, and those doing the rating give more weight to formants, which more accurately reflect vocal tract length because they correlate with the weight of the dog.⁷

Dogs are cognitive and complex. Their signaling behaviors are a reflection of these characteristics. Dogs use signals to convey information and to provoke the social situation to gain information. Dogs ask questions. Given their ability to identify specific dogs and contexts from barks, we should consider testing whether (and likely accept that) vocal signaling in dogs has rules (grammar) that is providing more information than is currently recognized. When we evaluate or classify canine behaviors we need to be acutely aware of such patterns and how inadequate our labels can be for conveying such complexity. 'Barking' can be involved in a number of behavioral conditions and client complaints. Addressing the underlying context is going to result in the most successful outcome.

'Nuisance' barking - even if the dog is behaving in a contextually appropriate manner - can cause the dog's death or relinquishment. Redress should be appropriate and humane. The single most appropriate intervention strategy is to acknowledge the dog and provide the information requested. This is easy to do if the dog is not distressed and the human is present: acknowledge the alert, check to see what the dog is barking at, and thank the dog. Reward the dog for being quiet. Barking in distress requires that the distress be addressed.

In neither of these situations are bark collars the best choice. Collars that make a spraying sound and/or spray the dog with citronella solution can startle and render more fearful susceptible dogs. Dogs can and will learn to bark below the level necessary to trigger the collar or to quickly empty the collar, but they will still not have their needs met. Because electric shock collars actually cause pain, they are likely to stop all behaviors associated with shock, including normal, social behaviors.

If shock and pain are profound, it is possible to quickly induce long-term potentiation (LTP), the molecular changes associated with hippocampal memory, which will lead to a strong aversion or phobia. The hippocampus is the primary region where fears and anxieties associated with specific stimuli originate, so a logical sequelae to a stressful, painful stimulus may be fear, phobia, or withdrawal. Less painful stimuli, over time, because of the nature of repeated reinforcement, ensures better, more numerous and more efficient connections between neurons.⁸ This is how fear and avoidance are learned. When stimulation continues, we know that activity dependent plasticity at synapses (e.g., learning) occurs in the lateral amygdale which is involved in learning contextual fear.⁹

Untoward, negative, long-term effects of training with shock have been demonstrated for German shepherd guard dogs.¹⁰ Dogs that were shocked in training showed a lower ear posture in free-walking, and more stress-related behaviors than did dogs who had not been shocked in training. These differences were also found when these dogs participated in obedience training and manwork. Physiological differences in the HPA axis were also found and were most profound when the person associated with the shock (the owner or handler) was present. The resulting conclusions were: 1) this type of training, in general, is stressful, 2) receiving shocks is painful for the dogs, and 3) the dogs make a context-specific association where the presence of the owner / handler and his / her commands provokes a stress response.¹⁰ Because the behaviors leading to the vocalization matter, we can expect shocks - even those delivered in response to barking - to be perceived by many dogs to be unpredictable. Laboratory research has shown that in normal dogs, such interventions cause stress and help produce pathological behaviors.¹¹

Early intervention, appropriate treatment and meeting the dog's needs are more humane solutions for dogs who are sharing with us information about how they perceive their world and its attendant discomforts.

References

1.  Molnar C, Kaplan F, Roy P, Pachet F, Pongracz P, Doka A, Miklosi A. Classification of dog barks: a machine learning approach. Anim Cogn 2008;11:389–400.

2.  Farago T, Pongracz P, Range F, Viranyi Z, Miklosi A. 'The bone is mine': affective and referential aspects of dog growls. Anim Behav 2010;79:917–925.

3.  Maros K, Pongracz P, Bardos G, Molnar C, Farago T, Miklosi A. Dogs can discriminate barks from different situations. Appl Anim Behav Sci 2008;114:159–167.

4.  Yin S, McCowan B. Barking in domestic dogs: context specificity and individual identification. Anim Behav 2004;68:343–355.

5.  Overall KL, Agulnick L, Kapes M, Dunham AE. Sonographic analysis of dog vocalization: a pilot study involving distressed and unaffected dogs (poster/abstract). American Veterinary Society of Animal Behavior (AVSAB) meeting, New Orleans, LA, July 1999.

6.  Pongracz P, Molnar C, Miklosi A. Acoustic parameters of dog barks carry emotional information for humans. Appl Anim Behav Sci 2006;100:228–240.

7.  Taylor AM, Reby D, McComb K. Human listeners attend to size information in domestic dog growls. J Acoust Soc Amer 2008;123:2903–2909.

8.  Wittenberg GM, Tsien JZ. An emerging molecular and cellular framework for memory processing by the hippocampus. Trends Neurosci 2002;25: 501–505.

9.  Schafe GE, Nader K, Blair HT, LeDoux JE. Memory conciliation of Pavlovian fear conditioning: a cellular and molecular perspective. Trends Neurosci 2001;24:540–546.

10. Schilder MBH, van der Borg JAM. Training dogs with help of the shock collar: short and long term behavioural effects. Appl Anim Behav Sci 2004;85:319–334.

11. Schalke E, Stichnoth J, Ott S, Jones-Baade R. Clinical signs caused by the use of electric training collars on dogs in everyday life situations. Appl Anim Behav Sci 2007;105:369–380.