When one Spanish word ends in a vowel and the next begins with one, do not expect the written space to create a pause. The vowels may stay fairly distinct, become more tightly linked, sound glide-like, or become more compressed depending on the phrase, stress, speaking rate, dialect and speaker.

Your eyes see mi | amigo. Your ears may hear one smooth stretch. Nothing was misspelled; the visual space was never a promise of a pause.

Use one simple rule: mark the vowel border, remove the imaginary pause, then let the audio decide what actually happened.

What does “merge” really mean?

For learners, merge is a useful description of the feeling: two written words stop sounding neatly separated. But it does not mean every pair of adjacent vowels becomes one identical sound.

Experimental phonetics on Spanish vowel sequences across word boundaries shows multiple realizations. Depending on the vowels and context, speakers can preserve more of the two-vowel sequence, create a diphthongal or glide-like transition, or produce stronger phonetic compression or blending. Recent work on Castilian Spanish explicitly argues against treating external vowel sandhi as one simple either/or process.

So instead of learning “two vowels = delete one,” learn three listening possibilities:

  • Two fairly clear vowels: the boundary is continuous, but both vowel qualities remain easy to hear.
  • A tighter glide-like transition: the two vowels run together more smoothly.
  • Stronger compression or blending: the written boundary becomes much harder to hear.

The exact acoustic result is the speaker’s job. Your job is to find the boundary and hear what happened there.

Why does the same vowel boundary sound different sometimes?

Because Spanish connected speech is not produced by a single “merge vowels” button. Research on Peninsular Spanish has found that factors such as vowel type, stress, articulation rate, prosodic position and word properties can influence what happens to adjacent vowels across a word boundary. Work on Mexican Spanish likewise shows more than one hiatus-resolution strategy across speakers and vowel combinations.

That gives learners an important safety rule:

Predict the location, not the exact collapse.

If you see a vowel at the end of one word touching a vowel at the beginning of the next, you have found a likely connected-speech pressure point. But do not decide in advance that one vowel “must disappear.”

Vowel Border Check: can you find the sound seam?

For each phrase, locate the written word boundary first. Then ask: Should I insert a pause here? Make your prediction before opening the explanation.

mi | amigo — should there be a tiny pause after mi?

No automatic pause. The /i/ at the end of mi meets the initial vowel of amigo. In connected speech, expect continuity across that border. The exact degree of glide or compression varies, so verify the real audio instead of inventing a fixed spelling.

no | es — should both words get separate little vowel beats?

Do not force separation just because you see a space. The /o/–/e/ sequence is a cross-word vowel contact. Say the phrase continuously, then compare with a speaker’s actual version.

la | amiga — should you delete one a?

Do not delete by rule. Identical or similar vowels can make the boundary especially hard to hear, but your safe target is a continuous phrase based on real audio—not “always erase one letter.”

fue | a Madrid — what should you notice first?

The e–a boundary. Mark it, remove the imagined pause, then listen for whether this speaker keeps the two vowels clearer or compresses the transition more strongly.

Is this sinalefa?

You will often see the term sinalefa or synalepha used for adjacent vowels across word boundaries behaving as one tighter syllabic unit. That label is useful, especially for recognizing the general phenomenon.

For pronunciation practice, though, the label is less important than the listening question: what did this speaker do with these two vowels in this phrase?

Acoustic research matters here because “one syllable” does not force one identical phonetic outcome. Two speakers—or the same speaker in careful and conversational speech—may realize the boundary differently.

The practice method: Mark → Stream → Verify → Say

Mark

Take a short phrase and insert a vertical bar at the written boundary: mi | amigo. The bar means “word boundary.” It does not mean “pause here.”

Stream

Say the whole phrase once without stopping at the bar. Do not deliberately delete either vowel. Your first goal is simply continuous airflow and timing.

Verify

Listen to a reliable phrase-level recording and classify what you hear:

  • two vowels still fairly clear;
  • a glide-like/tighter transition;
  • stronger compression or blending.

You do not need laboratory equipment. You are training a practical distinction: Did I imagine a pause that the speaker did not make?

Say

Repeat the whole phrase three times, copying the speaker’s timing rather than one isolated vowel. Then use the same vowel pattern in a different phrase.

If you can only reproduce mi amigo because you memorized one sound blob, you have learned one recording. If you can recognize another vowel-to-vowel boundary and handle it intelligently, you have learned a connected-speech habit.

Three habits that make Spanish sound more separated than it needs to

1. Pausing at every written space

This is not a grammar error, and careful separated speech may still be understandable. But mechanically inserting a tiny pause at every word boundary can make phrase rhythm unnaturally segmented and can train your ear to expect boundaries that conversational speech may not provide.

2. Deleting one vowel every time two vowels meet

This is an overgeneralization. Cross-word vowel sequences are variable. Stress, rate, vowel identity, dialect and speaker can all affect the realization.

3. Inventing a phonetic spelling and memorizing it

A homemade spelling may help you remember one recording, but it can also freeze one speaker’s realization into a fake universal rule. Mark the boundary instead. The actual audio is a better pronunciation authority than your keyboard.

A five-minute Space-Blind drill

Real dialogue is where the boundary becomes obvious

Careful word lists are useful for vocabulary, but they hide the exact problem this article is about: words touching other words. Once you understand the boundary, practise it inside whole lines. That fits a broader method of practising pronunciation from real dialogue.

On supported video pages with usable Spanish audio and subtitles, FunFluen can help you repeat a short line, move sentence by sentence, and reduce playback speed slightly while you inspect a difficult vowel boundary.

Loop one Spanish vowel boundary and test what you actually hear

The exact lesson from this article is not preloaded after the click, and playback or pronunciation aids do not certify one perfect pronunciation. Use the source audio as the reference.

Quick questions about Spanish vowel linking

Do two vowels always become one sound?

No. They can be produced continuously while remaining relatively distinct, or they can show tighter diphthongal, glide-like or more compressed realizations. The context and speaker matter.

Is every cross-word vowel connection called sinalefa?

Sinalefa is a common and useful label for adjacent vowels across words forming a tighter syllabic unit, especially in teaching and prosodic descriptions. For pronunciation practice, focus on the actual acoustic realization rather than assuming the label predicts one exact sound.

Should I delete a vowel to sound more natural?

No universal deletion rule is safe. First remove the artificial pause. Then copy the degree of connection you actually hear from a reliable speaker.

The space is on the page

Spanish speakers are not hiding vocabulary from you. Your eyes are simply giving you a visual boundary that your ears may not receive as a pause.

Mark → Stream → Verify → Say. Find the vowel border, keep the phrase moving, let the real audio tell you how tight the connection is, and practise the whole phrase.

The words are still there. The acoustic fence between them may be much smaller than the space on the page.

Sources