Try to pronounce “going to” slowly and cleanly in a casual conversation, and it sounds stiff, almost wrong. Say “gonna” instead, and that's what actually passes for correct, natural English. This isn't an exception to how English works — it's the default setting for fast speech, and there's a real mechanism behind why it happens, and why it's specifically hard to catch when Korean is your first language.
What Connected Speech Actually Does
The words don't change. What changes is how much of each one survives out loud.
Written
“What are you going to do?”
Spoken at normal speed
“Whaddaya gonna do?”
Written
“I want to eat.”
Spoken at normal speed
“I wanna eat.”
Written
“Did you eat yet?”
Spoken at normal speed
“Didja eat yet?”
Written
“I don't know.”
Spoken at normal speed
“I dunno.”
Written
“Give me a couple of minutes.”
Spoken at normal speed
“Gimme a couple minutes.”
Why This Specifically Trips Up Korean Speakers
Rhythm, not just words
English is stress-timed — some syllables compress almost to nothing to keep an even beat between the stressed ones. Korean runs closer to syllable-timed, where each syllable carries roughly equal weight. Neither is wrong. Connected speech is just English leaning hard on its own rhythm.
Korean doesn't have a matching process
Korean has no real equivalent to a whole word — “to,” “have,” “you” — getting swallowed into the sound next to it. There's no gap in your English knowledge to blame here. There's just no matching pattern in Korean to transfer from.
It's not slower for native speakers either
Native speakers can't cleanly slow down and produce “going to” in casual conversation without sounding stilted. The reduction isn't laziness — it's the default setting for fast, natural speech in any accent.
Words melt into each other, not just shrink
It's not only that individual words compress on their own — sounds at the boundary between two words merge into something new. “Meet you” drifts toward “meetchu.” “Did you” drifts toward “didja.” None of this is written anywhere; it only exists in the spoken version.
This is exactly what PopEar is for.
Textbook audio is recorded slow and clean on purpose — real speech never is, and no amount of grammar review changes that. PopEar's clips come straight from real shows at native speed, reductions and all, matched to your level, so your ear is training on the real thing instead of a cleaned-up stand-in.
Try PopEar FreeThe actual difference
The words in "gonna" and "wanna" were already familiar long before any of this. What your ear is missing is repetition in their real, reduced shape, enough times that the shape itself stops registering as mumbling.
Keep Learning
Why 'Did You Eat' Sounds Like 'Didjeet'
Not a contraction like gonna or wanna: two words melting into one new sound. PopEar's free iOS app trains your ear on real, fast, connected speech.
AI Workplace English: The New Office Vocabulary
Terms like 'prompt engineering' and 'human-in-the-loop' are now normal meeting talk. PopEar's free iOS app helps you follow along without missing a beat.
