Upspeak and Listener Judgments of Competence

Upspeak, also called the high-rising terminal or HRT, is a rising pitch contour that lands at the end of a declarative sentence. That distinction matters because the acoustic shape of upspeak is measurably different from the rise of a genuine yes-or-no question: the contour, its timing, and its peak all behave differently depending on whether a speaker is actually asking something or just ending a statement on a lift. Listeners hear upspeak and pick up on something more specific than a statement mistaken for a question, through a channel that has nothing to do with grammar.

The pattern has moved fast through British and American English, and younger speakers use it with particular frequency. That frequency matters on its own terms: listeners of every age now encounter rising declaratives often enough to have built strong, mostly unconscious expectations about what the pattern means. Those expectations do not form in a vacuum. Goupil et al., writing in Nature Communications in 2021, identified a cluster of acoustic features that listeners associate with both uncertainty and dishonesty in a speaker: rising intonation, reduced intensity at the onset of words, speech that runs slower and less evenly, and pitch that wanders more than usual. None of these features works alone. They show up together, as a signature, with upspeak sitting inside that signature.

That overlap is the hinge the rest of this piece turns on. If rising intonation were just one unrelated quirk among many, the competence penalty attached to it would be a curiosity. Instead it is one thread in a bundle that the human ear has apparently learned to treat as a single warning signal.

The prosodic signature listeners use to judge reliability

Listeners appear to carry an automatic mechanism for flagging unreliable speech, and it does not wait for permission from the conscious mind. Goupil et al. (2021) found that this prosodic judgment fires independently of what the listener actually knows about the content of the statement and independently of the listener's native language. The penalty is closer to a reflex than a verdict reached after careful listening, triggered by the acoustic cluster described above, and upspeak's rising contour sits close enough to that cluster to set it off.

The same research found that this effect reaches further than snap judgments. Words spoken with the "unreliable" prosodic profile get remembered differently than the same words spoken without it. That finding pushes this out of the realm of mild social friction and into something closer to information processing: the penalty is a difference in what actually gets encoded and retained, not a rating a listener might give if asked on a survey.

A speaker who uses a rising terminal to invite agreement, soften a claim, or keep a conversation collaborative is, from the inside, doing something entirely reasonable. A listener on the other end, operating a different and older piece of cognitive machinery, is registering a cue associated with uncertainty and possibly dishonesty, and that registration happens before any deliberate evaluation starts. The two parties are running two separate interpretive processes on the same sound rather than having a disagreement about tone, and one of those processes was never built to track intent. That mismatch is the actual subject of this piece: not whether upspeak is good or bad, but why a speaker's signal and a listener's reception can diverge so completely while both sides are certain they understood each other.

How listener age, speaker status, and context determine how severe the penalty is

None of this plays out the same way in every room. The acoustic trigger is consistent, but how harshly a listener penalizes it depends heavily on who is listening, who is speaking, and what the speaker's position already establishes before a word gets said.

Age turns out to be one of the sharpest dividing lines in the data. For listeners between 18 and 30, the presence of upspeak does not move competence ratings. For listeners between 60 and 72, upspeak users are rated significantly less competent. The acoustic signal has not changed between those two audiences; the decoding has. A speaker can walk out of a meeting with people their own age having registered as perfectly credible and walk into a meeting with an older panel sounding, to that panel, meaningfully less capable, without changing a single word of what was said.

Status complicates the picture further, and in a direction that cuts against the simple story. In business and academic meetings conducted in English in Hong Kong, meeting chairs, the people holding the most institutional power in the room, used rising intonation three to seven times more often than their subordinates. A 1991 study of a University of Texas sorority found a similar pattern: upspeak appeared most often in statements made by the more senior, more powerful members of the group. In both cases, rising intonation looks less like a marker of weakness and more like something closer to a prerogative of rank, a way an already-secure speaker can check in with a room without any risk of being read as unsure.

That reframes what the penalty is actually punishing. If upspeak signaled uncertainty in some fixed, acoustic sense, it would cost chairs and senior members just as much as it costs everyone else, but it doesn't. The people least in need of a credibility boost get to use the pattern freely, while the people with the most riding on being taken seriously, young speakers, newcomers, anyone without institutional standing already locked in, are the ones who pay for it. The penalty, in other words, falls hardest exactly where the stakes for sounding competent are highest.

The gender asymmetry: why the same rising intonation is heard differently depending on who produces it

If status can insulate a speaker from the penalty, gender appears to do the opposite, and the research on this point is specific enough to leave little room for argument. Roche et al., publishing in Frontiers in Psychology, found that rising intonation functions as a cue to low confidence when produced by either women or men, but listeners are more sensitive to that cue when a woman produces it. In a single-talker condition, a female voice using rising intonation was rated less confident than a male-pitched voice using the identical rising intonation. The acoustic input was the same. The rating was not.

That asymmetry was not fixed, though. In a multi-talker condition, where no single voice carried distinguishing qualities that separated it from the others regardless of gender, the gap disappeared. That detail matters because it tells us the bias is heuristic rather than hardwired: it depends on the listener having an easy gender cue available to latch onto. When that cue loses its salience, the penalty stops tracking gender.

The paradox buried inside this research is the sharpest part of the whole picture. Linguist Thomas J. Linneman has argued that the more successful a man becomes, the less likely he is to use HRT, while the more successful a woman becomes, the more likely she is to use it. Put those two patterns next to the Roche findings: the women most likely to reach for rising intonation, the high achievers, are reaching for the exact cue that listeners penalize most heavily when a woman produces it. It is a case of success correlating with heavier use of a signal that costs women more than it costs men for using it.

The stakes of that pattern are not confined to the seminar room. Research presented at the American Economic Association's 2024 conference examined market reactions to gendered speech patterns during corporate earnings calls and found that analysts and investors reacted negatively when female executives used rising intonation, while male executives speaking the same way faced no comparable penalty. The asymmetry is visible where it is hardest to dismiss as a lab artifact: in the pricing decisions of people moving real money.

Where the "upspeak always signals weakness" argument breaks down

The case made so far might suggest a simple rule: rising intonation reads as weak, so eliminate it. The evidence does not actually support that conclusion, and the linguistics literature is fairly blunt about why.

Fluency effects on competence judgments, upspeak included, are strongest when listeners have very little else to go on. Giving a listener more context, more content, and more time to form an impression shrinks the influence of a single prosodic cue. Upspeak also does real pragmatic work that has nothing to do with uncertainty: it can flag new information, keep a conversation moving, or check whether a listener is still following along. Researchers studying the indexical field of rising pitch have found that the same contour maps onto different social meanings for different listeners and in different communities. The "upspeak equals weakness" reading is a culturally specific decoding habit.

The ethical objection buried in all of this deserves serious consideration. Telling speakers to correct their intonation puts the burden of a listener's bias onto the person being judged, not the person doing the judging. The same critique applies to coaching aimed at vocal fry, regional accents, and other features disproportionately associated with women and younger speakers: correcting the speaker treats someone else's heuristic as the speaker's problem to fix.

That objection is correct as far as it goes, and it still does not change the practical situation a speaker faces in the room. The bias may belong to the listener, but the consequence, the job not offered, the pitch not funded, the earnings call that moves the stock the wrong way, lands on the speaker regardless of who is at fault. Awareness of that gap is a form of leverage rather than an admission that the bias is fair, because a speaker who understands exactly where the gap sits is in a far stronger position to decide what to do about it than one who has never been told it exists.

What it means to have control over your intonation rather than just suppressing one pattern

The goal that follows from all of this is not deleting the rising terminal from a speaker's repertoire but building enough range that a rise occurs when the speaker means to invite engagement and a fall occurs when the speaker means to close a point with certainty. Treating upspeak as a defect to be stamped out produces speakers who suppress a pattern they never fully understood. Treating it as one tool among several produces speakers who can pick it up or set it down depending on what a given sentence actually needs, which is a considerably more advanced skill than simply staying quiet about it.

A few concrete habits move a speaker from the first category to the second. Recording a short stretch of unscripted speech and listening back for sentences that rise without any real reason is one of the fastest ways to find the pattern in the wild, since most people cannot hear their own intonation in real time. Practicing a slight downward inflection on the final word of a declarative sentence trains the ear and the voice together. Picturing the period at the end of a sentence as a physical instruction, something that closes the pitch rather than leaving it hanging open, gives the brain a concrete target instead of a vague instruction to "sound more confident."

The acoustic fix tends to follow from the underlying habit of seeking approval. Assertive communication, making a claim and letting it stand without soliciting immediate agreement, builds the kind of settled confidence that reduces the unconscious, approval-seeking use of rising intonation. Train the habit of stating things and the pitch tends to follow; train the pitch alone and the habit often snaps back the moment attention drifts.

Because upspeak is strongly associated with younger speakers, conscious downward inflection on key statements is also one of the fastest adjustments available to anyone trying to project more gravitas in a high-stakes moment: a presentation, an interview, a first introduction where the room has not yet decided what to make of the speaker. None of this requires an overhaul. A short daily habit, recording a single spoken answer, listening back, naming what the pitch did at the end of each sentence, builds the tight feedback loop that turns this from something a speaker knows about in the abstract into something a speaker can actually hear and adjust in real time. The gap between intent and perception does not close by reading about intonation. It closes through repetition.

Why the generational perception gap makes intonation awareness more urgent, not less

Upspeak is common among younger speakers partly because it functions as a marker of identity and authenticity within that group, which is a legitimate thing for a speech pattern to do. The complication is that the people holding the most consequential evaluative power, hiring managers, senior colleagues, investors, interview panels, skew toward the older end of the age range the research identifies as most sensitive to the competence penalty. The same rising terminal that reads as neutral among peers can read as a liability in front of exactly the audience whose judgment carries the most weight.

None of this means a young speaker should sound like someone else; it is an argument for understanding the decoding process well enough to choose, deliberately, which signal gets sent in which room. The data on age groups already shows that the same speaker, using the exact same intonation, can be rated as competent by one audience and less competent by another without changing anything other than who is listening. Intonation awareness is what lets a speaker move between those two audiences without having to pick one identity and sacrifice the other.

Code-switching across context is not a new demand; people have always adjusted register, vocabulary, and tone depending on who they're talking to. What is new is the degree to which the acoustic gap between what a speaker means and what a listener hears can now be measured, named, and practiced against, rather than treated as an invisible mismatch nobody can do anything about. The speakers who get ahead of that gap, who can call up falling intonation when the moment calls for authority and rising intonation when the moment calls for openness, are the ones who end up sounding both confident and human, rather than boxed into one read by a listener running an interpretive process the speaker never knew was active.