A dog whistle is a small instrument that produces sound at frequencies too high for most humans to hear but well within a dog’s hearing range, and the same term has become one of the most debated concepts in political communication, describing coded language that carries a hidden meaning for a specific audience while sounding innocent to everyone else. The two uses share more than a name: both rely on the principle that different listeners can hear the same signal in fundamentally different ways. The literal whistle has a surprisingly rich scientific backstory, and the metaphorical one has grown into a subject of serious academic research spanning political science, linguistics, law, and even artificial intelligence.
The Whistle Francis Galton Built
The device most people picture when they hear “dog whistle” traces back to the Victorian polymath Francis Galton. In the 1870s, Galton hypothesized that the upper limit of hearing for high-pitched tones might vary from species to species, and he built a small adjustable whistle to test this idea.1PubMed. The Galton whistle and discovery of presbycusis The Galton whistle, as it came to be known, could be tuned to produce sounds well above the range of human perception. By observing which animals responded to which pitches, Galton confirmed what dog owners had long suspected: dogs can hear sounds that are inaudible to us. His experiments also led him to notice that human hearing deteriorates with age, contributing to the early understanding of presbycusis, or age-related hearing loss.
Modern dog whistles work on the same principle. Most operate in the range of roughly 23,000 to 54,000 Hz, though adjustable models let the user dial in a specific pitch. For comparison, the upper limit of adult human hearing sits around 20,000 Hz and typically drops lower with age. A dog whistle set to, say, 35,000 Hz produces a sound that a nearby person might perceive as a faint hiss or nothing at all, while a dog across a field perks up immediately.
Why Dogs Hear What We Cannot
Dogs have evolved an auditory system that is sensitive to a much broader frequency range than ours, extending up to about 65,000 Hz in some breeds. This is partly a legacy of their ancestry as predators whose prey, like small rodents, communicate with high-frequency calls. Their ears are also remarkably good at pinpointing where sound is coming from. Research on canine sound localization has found that dogs can distinguish the direction of a sound source separated by as little as about eight degrees on average, with some individuals performing even better.2PubMed Central. Sound Localization Ability in Dogs This precision helps explain why a whistle blown from hundreds of meters away can still draw a working sheepdog’s attention in the right direction.
Breed matters, though not as dramatically as you might expect. Dogs with erect, forward-facing ears, like German Shepherds, tend to capture high-frequency sound slightly more efficiently than breeds with large, floppy ears. But even a Basset Hound hears frequencies far above what any human can manage. Age plays a bigger role: just as in humans, hearing sensitivity in dogs declines over time, particularly in the higher frequencies. An older dog may stop responding to a whistle not because it has lost interest but because the pitch has crept outside its diminished range.
Do Ultrasonic Devices Actually Work on Dogs?
Dog whistles sold for training are one thing, but the market is full of ultrasonic gadgets marketed for other purposes: deterring aggressive dogs, stopping nuisance barking, repelling strays. The evidence for these devices is a lot weaker than the packaging suggests. A study testing multiple ultrasonic units on dogs found that no single device repelled all dogs at any frequency or sound pressure level tested. One commercial unit produced an immediate reaction in most dogs tested, with some responding aversively, but the reactions varied widely from dog to dog.3Applied Animal Behaviour Science. Aversive responses of dogs to ultrasonic, sonic and flashing light units Some dogs showed clear signs of distress, others merely pricked their ears, and a few seemed unbothered.
The inconsistency comes down to individual variation in hearing sensitivity, temperament, and habituation. A bold, confident dog may register an ultrasonic burst and decide it is not a threat. A more anxious dog may startle but then learn to ignore the noise after repeated exposure. For anyone considering an ultrasonic device to manage an aggressive stray or a neighbor’s barking dog, the honest takeaway is that these are unreliable tools. They are not harmless either: a sound that causes genuine distress to one dog while failing to deter the target animal raises obvious welfare concerns.
The same problem extends to wildlife. Ultrasonic whistles have been attached to vehicles and trains in an effort to scare deer and other animals off roads and tracks, but tests on species including white-tailed deer, black-tailed deer, dingoes, and birds have repeatedly found these devices to be ineffective.4Elsevier (Transportation Research Part D: Transport and Environment). Effectiveness of an acoustic wildlife warning device using natural calls to reduce the risk of train collisions with animals If an ultrasonic pulse will not reliably move a deer off a railway, it is worth being skeptical of consumer products that claim to repel dogs from your garden.
How an Acoustic Gadget Became a Political Metaphor
Sometime in the late twentieth century, commentators began borrowing the term “dog whistle” to describe a very different kind of signal: political language designed to be heard one way by a target audience and another way, or not at all, by everyone else. The analogy is intuitive. Just as a literal dog whistle transmits a message that only certain ears can decode, a rhetorical dog whistle embeds a meaning that only certain listeners pick up on, while sounding perfectly ordinary to the rest.
The concept has been studied most extensively in the context of race and American politics. Scholars have documented how racially coded language that appeals to racial bias without overt bigotry has a long history in the United States, with politicians deliberately activating latent racial attitudes in voters without ever explicitly mentioning race.5The American Journal of Economics and Sociology. Cracking the Racial Code: Black Threat, White Rights and the Lexicon of American Politics Phrases about “law and order,” “welfare reform,” “inner cities,” or “states’ rights” can serve as vehicles for racial messaging while maintaining enough surface-level innocence to avoid direct accusations of racism. This is the mechanism that makes the metaphor so apt: plausible deniability is baked into the design.
How Coded Political Language Actually Functions
The mechanics of a political dog whistle rely on a gap between what a phrase means on its surface and what it signals to a subset of listeners who share specific cultural knowledge or political attitudes. A politician who speaks about “protecting our neighborhoods” is making a statement that sounds unobjectionable to most people. But for an audience already primed by decades of racially coded discourse, the phrase can activate specific associations about who belongs in those neighborhoods and who represents a threat.
Experimental evidence backs this up. A study examining racial dog whistles in Sweden found that coded terms can help politicians gain support from sympathetic in-groups while retaining backing from out-groups who do not perceive the hidden meaning. When researchers compared reactions to an explicit phrase like “immigrant gang” versus a coded alternative like “suburban gang,” the coded version generated more political support overall.6Journalism & Mass Communication Quarterly. Coded Appeals and Political Gains: Exploring the Impact of Racial Dogwhistles on Political Support The overt term alienated some voters, while the coded one slipped past their defenses. This is the strategic advantage of dog-whistling: it lets a speaker appeal to prejudice without paying the social cost of open prejudice.
Research on American politics has found a similar dynamic with populist rhetoric. Anti-elite language on the political right has been shown to function as a kind of dog whistle by bundling economic grievance with ethno-nationalist sentiment, so that listeners who share those attitudes hear a reinforcing message while others hear only generic anti-establishment talk.7Social Forces. Populism as Dog-Whistle Politics: Anti-Elite Discourse and Sentiments Toward Minority Groups The coded nature of the communication means that calling it out directly is difficult, because the speaker can always point to the innocent surface meaning.
Dog Whistles Are Not Just About Race
While racial coding has attracted the most scholarly attention, the mechanism is not limited to race or even to politics in the traditional sense. Researchers have argued that dog-whistling should not be reduced to racism or verbal communication tactics alone but can be combined with various worldviews and can encompass verbal and non-verbal communicative means that surreptitiously signal to a specific subgroup.8The British Journal of Politics and International Relations. Dog-whistling and democracy A gesture, an image, a musical choice, even a specific font or color scheme on campaign material can function as a dog whistle if the right audience recognizes it and the broader public does not.
One of the more striking examples comes from advertising rather than politics. In the 1990s, the car company Subaru discovered that lesbians were a disproportionately loyal customer base but recognized that openly marketing to gay women could alienate other buyers in a less tolerant cultural climate. The company began embedding references in its ads that lesbian audiences would immediately recognize while the general audience remained unaware. Taglines were crafted with deliberate double meanings, and imagery carried signals visible only to those looking for them.9Analisa. The Definition and Typological Model of a Dogwhistle The strategy worked: Subaru built a fiercely loyal following within the lesbian community while maintaining broad mainstream appeal, eventually building enough cultural capital to move toward openly inclusive messaging. This is a case where dog-whistling served an inclusive rather than exclusionary purpose, complicating any assumption that coded communication is inherently harmful.
Tracking Dog Whistles in Digital Spaces
The internet has made dog whistles both more potent and harder to pin down. Online communities can develop coded vocabularies at remarkable speed, cycling through new terms as old ones get flagged and lose their plausible deniability. A phrase that serves as a dog whistle in one forum may be completely unknown to moderators, journalists, or researchers monitoring the same platform for extremist content.
Computational linguists have begun developing tools to keep pace. One research team created a method for tracking how dog-whistle terms evolve over time in online discussion forums by modeling how the contextual meaning of words shifts within specific communities. The approach combines computational semantics with survey methods, mapping in-group and out-group interpretations of the same terms to reveal hidden layers of meaning.10Journal of Computational Social Science. Tracking dogwhistles online and through time using distributional semantics This kind of work is inherently playing catch-up, since the communities generating coded language are constantly adapting, but it offers a more systematic alternative to relying on individual journalists or activists to spot new terms.
A separate large-scale project assembled a glossary of over 300 dog-whistle terms, complete with contextual information and examples, and analyzed their usage in historical speeches by American politicians. The researchers also tested whether a large language model could identify dog whistles and their intended meanings, finding that the AI’s ability varied dramatically depending on the type of dog whistle and the group being targeted.11Association for Computational Linguistics. From Dogwhistles to Bullhorns: Unveiling Coded Rhetoric with Language Models Some well-established coded phrases were recognized easily; newer or more niche terms stumped the model entirely. The gap is worth noting because it mirrors the human challenge: recognizing a dog whistle requires the cultural context to decode it, and that context is not evenly distributed.
The Legal Problem With Hidden Meanings
Courts face a genuinely difficult challenge when coded language intersects with laws against hate speech or incitement. If a speaker uses a phrase that carries a violent or hateful meaning for a specific audience but sounds innocuous to everyone else, how should the law treat it? This tension has been examined through the lens of European Court of Human Rights case law, where scholars have explored how courts can assess speech that operates on multiple semantic levels simultaneously, balancing free expression protections against the risk of disguised incitement to hatred or violence.12Politics & Security. Dog whistles and cat senses – hidden extremism and the framework of legal oversight in European Court of Human Rights case law
The core difficulty is evidentiary. Proving that a speaker intended a coded meaning requires establishing what the audience understood, which is hard to demonstrate in a courtroom. A politician who speaks about “traditional values” in a context that a specific audience reads as anti-immigrant messaging can always point to the phrase’s literal meaning. Courts that try to go beyond the surface text risk appearing to police thought rather than speech, while courts that take language only at face value risk letting genuinely harmful incitement pass unchallenged. Digital environments compound this problem, since memes, irony, and layered references make it even harder to establish a single “meaning” for a given piece of communication.
Strategies for Countering Coded Language
If dog whistles work because they hide a message inside an innocent surface, one intuitive counterstrategy is to simply expose the hidden meaning: call it out, make the coded signal legible to everyone. But research on this has yielded mixed results. Calling out a dog whistle can sometimes backfire by amplifying the message, drawing attention to it, or making the accuser look paranoid to listeners who genuinely did not hear the coded meaning.
A different approach that has attracted scholarly interest is reframing. Instead of trying to decode and expose the hidden message, this strategy involves shifting the terms of the conversation entirely, redirecting the discussion to exclude the coded topic and introduce different ground. The idea is to change the conversation rather than win an argument about what a particular phrase really means.13Manuscrito. Code words and (re)framing This avoids the trap of trying to prove intent, which is almost always contested, and instead takes control of what gets discussed. Whether this works in the fast-moving, algorithmically sorted landscape of social media is another question. Platforms that amplify emotionally charged and binary narratives tend to reward the kind of content that dog whistles thrive on, making reframing harder to sustain at scale.14Journal of Sociocybernetics. Algorithmic Hegemony: AI, Political Elites, and the Reinvention of Electoral Influence
When the Whistle Becomes a Bullhorn
One of the more interesting dynamics researchers have observed is what happens when a dog whistle loses its coded status. Terms that once operated below the radar of mainstream discourse can get exposed, debated, and widely recognized, at which point they stop functioning as dog whistles and become something more like open signals. The glossary of over 300 dog-whistle terms compiled by computational linguists reflects this lifecycle: some entries are archaic phrases whose coded meaning is now widely understood, while others are so recent and niche that most people outside a specific online community would not recognize them at all.15Association for Computational Linguistics. From Dogwhistles to Bullhorns: Unveiling Coded Rhetoric with Language Models
This lifecycle creates a kind of arms race. Once a coded term enters mainstream awareness, its usefulness as a covert signal evaporates. Speakers who relied on that term either shift to new language or, in some cases, drop the pretense entirely and say openly what they used to encode. The shift from dog whistle to bullhorn, as some researchers describe it, is not always gradual. Political moments of crisis or polarization can accelerate the process, making previously unspoken ideas suddenly speakable. Understanding this cycle matters because it means that any snapshot of “current dog whistles” is inherently temporary. The vocabulary is always in flux, driven by the same social pressures that created the need for coded language in the first place.
For the literal dog whistle, the physics has not changed since Galton’s day. For the metaphorical one, the science of studying it is still catching up with how fast the language moves. Both versions continue to demonstrate the same underlying principle: a signal’s power depends as much on who is listening as on what is being said.