A text can be grammatically correct, stylistically polished, logically organized, and genuinely intelligent — and still become surprisingly weak when someone stands up and reads it aloud.
This is not necessarily a problem with the speaker.
It may be a problem with the text.
A page and a human voice do not deliver language in the same way. A reader can stop. A listener cannot stop the speaker mentally while the next sentence continues to arrive. A reader can move backwards, reread a difficult clause, inspect the previous paragraph, notice punctuation, and reconstruct an argument at their own pace.
A listener receives language as an event unfolding in time.
That difference changes almost everything.
Sentence length matters differently. So do rhythm, pauses, repetition, emphasis, transitions, vocabulary, information density, structure, audience, delivery, and memory.
A speech, therefore, is not simply a written text delivered aloud.
Writing for the eye and writing for the ear are different communicative constructions.
The Reader Has a Page. The Listener Has a Moment.
Consider this sentence:
The implementation of the proposed changes, while potentially beneficial in terms of long-term organizational efficiency, requires careful consideration of the financial, operational, and personnel-related consequences that may arise during the transitional period.
There is nothing grammatically wrong with it.
On a page, it may work perfectly well in the right professional context. The reader can see its structure. The commas help. The sentence remains physically available while the reader processes the relationships between implementation, efficiency, consequences, and transitional period.
Now imagine hearing it once in a conference room.
By the time the listener reaches personnel-related consequences, they may still be mentally organizing long-term organizational efficiency. But the sentence has continued moving while their processing has not.
A version designed for speech might be:
The changes could make the organization more efficient in the long term. But the transition will have consequences. We need to look carefully at the cost, the operational impact, and the effect on our people.
The second version is not necessarily “better English.”
It is better designed for this channel.
The information has been distributed across time. The listener receives one manageable unit and then another. The contrast is audible. The list has a clear entrance. The final element — our people — can receive natural emphasis.
The difference is architectural.
Written Language Can Wait. Spoken Language Cannot.
A written sentence exists in space.
A spoken sentence exists in time.
When reading, you control the speed of intake. Difficult information can slow you down. Familiar information can be scanned quickly. A paragraph remains visible while you interpret it.
Speech normally gives the listener much less control.
The speaker chooses the pace. The next phrase arrives whether or not the previous one has been fully processed. Unless the speech is recorded and replayed, its acoustic form disappears immediately.
This means that speechwriting must account for a simple but powerful fact:
the audience is processing while the language is still arriving.
Information density is therefore not merely a stylistic question. It is a processing question.
A sentence may contain exactly the amount of information a reader can handle and still contain too much information for a listener to process in real time.
This does not mean speeches must consist of primitive vocabulary and tiny sentences.
It means complexity has to be managed differently.
Writing and Speech Have Different Communication Chains
Ordinary writing can be represented, very roughly, as:
writer → text → reader
Speech introduces another set of transformations:
speechwriter → written speech → speaker → spoken language → listener
Each arrow matters.
The speechwriter may intend one thing.
The manuscript may encode it imperfectly.
The speaker may interpret the manuscript differently.
Pronunciation, rhythm, emphasis, speed, pauses, facial expression, and voice may change its effect.
The listener then receives not the manuscript, but the performance.
And what the listener understands is not necessarily identical to what the speaker said.
A more complete model is therefore:
what the speaker intends
→ what is written
→ what is actually said
→ what the audience hears
→ what the audience understands
→ what the audience remembers
Those are not six names for the same thing.
They are six stages at which communication can change.
This is one reason professional speechwriting involves much more than correcting grammar or making sentences elegant. It belongs to a broader understanding of language as a system of decisions: decisions about meaning, audience, sequence, form, context, and effect.
The Speaker Is Part of the Text
A written article can survive considerable distance from its author.
A speech cannot escape its speaker so easily.
Words that sound natural from one person may sound artificial from another.
Imagine a speechwriter produces:
We stand today at an inflection point that compels us to reimagine the institutional architecture through which our collective aspirations may be realized.
The sentence may fit a particular speaker, audience, institution, and occasion.
But give it to a manager who normally speaks in direct, concrete English, and the audience may hear not authority but borrowed language.
The problem is not merely vocabulary.
The text and the speaker do not belong to the same voice.
A more credible version might be:
We have reached a point where the old system is no longer enough. If we want different results, we need to change how we work.
This is why speechwriting cannot be separated completely from delivery.
The writer is not composing only semantic content. The writer is creating language that another human being must physically produce and socially inhabit.
The speaker has lungs, habits, an accent, a tempo, a personality, a status, a relationship with the audience, and a particular reason for speaking.
All of these become part of the communication.
Punctuation Is Visible. Pauses Must Be Heard.
On a page, structure has visual support.
Paragraphs separate ideas. Commas indicate relationships. Colons announce explanations or lists. Dashes interrupt or redirect. Headings show hierarchy.
Readers can literally see that one section has ended and another has begun.
Listeners do not see paragraph breaks.
Speech therefore needs audible architecture.
That architecture can be created through pauses, changes of pace, repetition, contrast, explicit transitions, strategic emphasis, parallel structures, short orienting sentences, and changes in delivery that signal a new stage of the argument.
This explains why repetition, sometimes unnecessary in tightly edited prose, can become functional in speech.
A reader can look back.
A listener may need the speaker to bring the idea back.
Repetition Is Not Always Redundancy
Written editors are often trained to remove repetition.
Frequently, they should.
But spoken communication changes the calculation.
Consider:
Our first problem is access. The second is quality. The third is continuity.
The repeated structure is doing work.
It tells the listener where they are.
Now compare:
The principal issues concern access, followed by questions of quality and, finally, the challenge of maintaining continuity.
The second sentence is compact and perfectly readable.
The first may be easier to follow in a room.
Speech sometimes needs what writing can afford to omit.
The key question is not simply:
Have I already said this?
It is:
Does the listener need to hear this again, in this form, at this moment, to follow the argument?
That is a different editorial decision.
Transitions Have to Survive Without Typography
Imagine a written article with three headings:
The problem
Why it happens
What should change
The reader sees the structure before reading a single sentence underneath them.
Now remove those headings and speak the article.
Where did the structure go?
Unless the speaker rebuilds it through language and delivery, much of it disappears.
A speech may therefore need sentences such as:
So what is actually going wrong?
or:
That explains the problem. Now we need to ask why it happens.
or simply:
There is a second issue.
These sentences may look almost too obvious on the page.
That can be exactly why they work aloud.
A listener needs orientation before becoming lost, not after.
Vocabulary Has an Acoustic Life
A word on a page is not only a word in a voice.
Long, unfamiliar, abstract, easily confused, or difficult-to-pronounce words create different risks when spoken.
This does not mean a speechwriter should avoid sophisticated vocabulary.
It means vocabulary must be evaluated in another dimension.
Do not ask only:
Is this word precise?
Ask:
Will the audience recognize it immediately when they hear it?
And:
Can this particular speaker say it naturally?
A technical term may be essential. If so, the speech may need to prepare the listener for it, explain it, repeat it, contrast it with something familiar, or place it in a sentence where its meaning can be recovered.
Written language can lean on spelling.
Spoken language cannot.
This distinction becomes especially important in multilingual and international communication. Someone may know a word perfectly on the page yet fail to recognize it immediately in connected speech.
That is one reason work on international meetings and presentations cannot be reduced to knowing the correct professional vocabulary. The language must also survive real-time communication.
Information Density Is Not the Same as Intellectual Depth
One of the most damaging assumptions in professional communication is that simplifying the processing path means simplifying the thought.
It does not.
A difficult idea can be expressed through a sequence of clear sentences.
A sophisticated argument does not require every sentence to carry five layers of information simultaneously.
Compare:
Although technological adoption is frequently discussed primarily in terms of infrastructure and financial investment, its success is ultimately dependent upon the extent to which individuals understand, trust, and integrate the technology into their everyday professional decisions.
Now reorganize the same idea:
We often talk about new technology as an investment problem. Do we have the infrastructure? Do we have the money? Those questions matter. But they are not enough. Technology succeeds only when people understand it, trust it, and actually use it in their everyday work.
The intellectual claim has not been removed.
Its route through the listener’s mind has changed.
This distinction is essential:
clarity is not the opposite of depth.
Sometimes clarity is what allows depth to arrive.
The same principle matters in professional writing, although the balance changes when the final recipient is a reader rather than a listener.
Rhythm Is Part of Meaning
Speech has a physical dimension that writing can partially hide.
Some sentences are technically pronounceable but exhausting to say.
Others create emphasis almost automatically.
Compare:
We need a policy that is clear, realistic, enforceable, transparent, and sufficiently flexible to respond to changing circumstances.
with:
The policy has to be clear.
It has to be realistic.
It has to be enforceable.
And it has to be flexible enough to survive change.
The second version takes more space.
But speech does not pay for space in the same currency as writing.
It pays in time, attention, breath, anticipation, and memory.
Rhythm helps organize those resources.
Short and long sentences can alternate. Parallel structures can create expectation. A pause can separate an important conclusion from the reasoning that produced it.
Sometimes a very short sentence carries more force precisely because the preceding language has prepared the silence around it.
A Pause Is Not Empty
On a page, silence is difficult to represent.
In a room, silence can be part of the message.
A pause can allow the audience to finish processing an idea, anticipate what comes next, recognize a transition, react, reconsider, or simply remember.
A manuscript that contains no room for pauses may force the speaker into continuous verbal delivery.
That creates a strange paradox:
the speech contains more words but communicates less.
The professional question is therefore not only:
What should be said here?
Sometimes it is:
Should anything be said here yet?
Silence itself can be a communicative decision. The broader relationship between language and what remains unsaid is explored in When Silence Speaks: The Language of What You Don’t Say.
Speechwriting Is Language as Decisions
Speechwriting makes one fundamental property of language especially visible: communication consists of decisions.
The writer has to decide:
What needs to be said?
What can be omitted?
What must come first?
What does the audience already know?
What might they misunderstand?
Which word is precise enough for the idea but natural enough for the speaker?
Where does the listener need repetition?
Where should the argument accelerate?
Where should it stop?
What should the audience understand immediately?
What should they remember ten minutes later?
And perhaps most importantly:
What effect is this language supposed to produce in this audience, through this speaker, in this situation?
Sentence length, rhythm, repetition, vocabulary, transitions, and pauses are not isolated tricks.
They are consequences of decisions about communication.
The technique follows the purpose.
There Is No Universal “Speech Style”
“Write short sentences” is useful advice until it becomes a rule.
A ceremonial speech is not a technical briefing.
A university lecture is not a wedding toast.
A political address is not a project update.
A five-minute conference introduction is not a forty-minute keynote.
A speaker addressing specialists can assume knowledge that would confuse a general audience. A speaker working in a second language may need different syntactic and lexical choices from a native speaker communicating the same conceptual content. A teleprompter creates different conditions from notes. A live audience creates different possibilities from a recorded video.
Speechwriting is therefore situational.
The right question is not:
What does a good speech look like?
It is:
What must this speech do, for these listeners, through this speaker, under these conditions?
There is no serious universal formula because the variables are not universal.
The Audience Does Not Receive Your Intention
Writers and speakers often evaluate communication from the wrong end of the chain.
They ask:
Did I express what I wanted to say?
That matters.
But communication does not end with intention.
Suppose the intended message is:
We need to change, but the change will be gradual and people will be supported.
The manuscript becomes:
The organization will commence a phased structural transformation designed to optimize operational effectiveness while implementing appropriate support mechanisms.
The speaker reads it quickly.
Some listeners hear structural transformation and assume layoffs.
Others miss phased.
Others do not know what support mechanisms means.
The speaker intended reassurance.
The audience heard threat.
Which message was communicated?
Professional language work has to examine the entire chain:
intention → formulation → delivery → perception → interpretation → memory
Success at the first stage does not guarantee success at the last.
Memory Changes the Design
After reading an article, you can reopen it.
After hearing a speech, what remains?
Usually not the complete wording.
Listeners retain fragments: a central idea, a contrast, a story, a phrase, a conclusion, an emotional impression — perhaps several main points if those points were structured clearly enough.
This should influence speechwriting from the beginning.
If everything is emphasized, nothing is emphasized.
If every sentence introduces new information, the listener has no time to consolidate the previous information.
If the central idea appears only once, hidden inside a dense paragraph, the speech may technically contain its message without successfully communicating it.
A speechwriter therefore works not only with sentences but with anticipated memory.
What deserves to survive?
And what linguistic structure gives it the best chance?
The Manuscript Is Not the Final Product
This may be the most important professional distinction.
For an article, the written text is close to the final communicative object.
For a speech, the manuscript is an intermediate object.
The final product is the event:
a particular person saying particular words, in a particular voice, at a particular pace, to particular people, in a particular situation.
That is why a speech should eventually be tested aloud.
Not merely proofread.
Spoken.
A sentence that looks excellent may reveal an awkward rhythm.
A transition that seemed obvious may disappear.
A paragraph may take much longer than expected.
A supposedly powerful phrase may sound theatrical in the speaker’s actual voice.
A difficult name may disrupt the speaker’s pace.
A joke may require a pause that the manuscript never anticipated.
The voice discovers problems the eye cannot.
Editing a Speech Means Listening Forward
When editing prose, we often ask whether a sentence connects logically with what came before.
Speech requires another question:
Can the listener follow where this is going while it is happening?
Listeners do not need to know the conclusion in advance. But they need enough structure to build a continuously updated mental model.
That is why good speech transitions often point forward:
Here is the problem.
There are two reasons for this.
The first is obvious. The second is more important.
Before we discuss the solution, we need to understand one thing.
These expressions are not intellectually impressive by themselves.
They are navigation.
And navigation becomes especially valuable when the road disappears behind you as quickly as you travel along it.
Writing Your Own Speech and Writing for Someone Else Are Different Tasks
A person writing their own speech already knows their intentions, habits, experiences, and preferred vocabulary.
That is an advantage.
But it can also create blindness. What feels obvious internally may never become explicit in the manuscript.
A professional speechwriter or editor faces a different problem.
They must understand not only the topic but the speaker.
What does this person actually sound like?
What can they say convincingly?
Which phrases belong to them?
Which arguments require more explanation?
What level of formality fits their role?
Where are they likely to accelerate?
Which words may create pronunciation problems?
How will the audience perceive this person before the first sentence is spoken?
Professional speechwriting therefore involves a form of linguistic modelling.
The writer is not trying to display their own style.
They are building language that must survive another person’s voice.
A Practical Test: Read It With Your Ears
When reviewing a speech, do not ask only whether the text is correct.
Ask whether you can understand each important sentence the first time you hear it.
Ask whether the important information arrives before attention becomes overloaded.
Ask whether you can hear where one idea ends and another begins.
Ask whether the speaker has room to breathe.
Ask whether repetition helps orientation or merely repeats.
Ask whether transitions remain clear without headings.
Ask whether the vocabulary belongs to this speaker.
Ask whether the audience will recognize important terms by sound.
Ask where the speaker should slow down.
Ask where the audience needs a pause.
And ask:
What is the one idea that should remain after the exact words disappear?
Then read the speech aloud.
Not silently while imagining your voice.
Actually aloud.
The difference is diagnostic.
Good Writing and Good Speech Overlap — But They Are Not Identical
None of this means that strong writing becomes irrelevant to speech.
Precision matters.
Structure matters.
Grammar matters.
Vocabulary matters.
Logic matters.
Style matters.
But speech adds another layer of constraints and possibilities.
The text must pass through a human voice.
The voice must pass through time.
The audience must process the message while it is happening.
And the message must survive the distance between what was intended and what is finally understood.
That is why writing for the ear is not simply “simpler writing.”
It is writing designed for a different path.
A good written text asks:
Does this work when someone reads it?
A good speech asks something more demanding:
Will this thought survive the journey from intention, to manuscript, to voice, to hearing, to understanding, to memory?
That is the real work.
Professional Language Work Beyond the Classroom
Language learning remains at the core of Levitin Language School, together with Academic & Professional Subjects and the Language + Subject approach.
But language is also professional work.
Depending on the task and the competence of the relevant specialist, this work can include editing, professional writing, speechwriting, presentation language, and adapting complex ideas for a particular audience.
The principle remains the same: language is not merely a collection of words and rules.
It is a sequence of decisions made for a purpose.
If you are preparing a speech, presentation, professional text, or other important communication, the useful question is not only whether the English is correct.
The more important question may be:
Will the language do what you need it to do when it reaches another person?
Tymur Levitin
Founder and Director, Levitin Language School
© 2026 Tymur Levitin / Levitin Language School. All rights reserved.
This material is part of the Levitin Language School educational and professional language ecosystem. Reproduction, adaptation, or republication without permission is prohibited.
