osoparavos.com

— Adam Donaldson Powell

Beyond Fluency: Thinking Poetically in a Second Language

Introduction

For most language learners, fluency is imagined as the destination. Vocabulary grows, grammatical errors diminish, pronunciation improves, and eventually conversation becomes effortless. Yet many advanced multilinguals discover an unexpected paradox. At precisely the moment they become capable of discussing philosophy, politics, or science in another language, they find themselves unable to write a convincing poem.

The obstacle is no longer linguistic competence in any conventional sense. It is aesthetic competence.

A poet does not merely communicate ideas. A poet chooses among thousands of equally grammatical possibilities, selecting words not only for what they denote but for what they evoke, where they belong in literary history, how they resonate with surrounding sounds, how they interact with rhythm, silence, metaphor, memory, and expectation.

Native speakers acquire these intuitions almost invisibly over decades of immersion. The bilingual poet must acquire them consciously.

This essay series on writing poetry in a foreign language is essentially a book that has never before been written in quite this way; and the “book” explores that journey in ways that I hope are beneficial.

Its central argument is simple:

Writing poetry in a second language is not primarily a matter of acquiring more vocabulary or mastering more grammar. It is the gradual development of a second aesthetic intuition.

That intuition cannot be memorized from textbooks. It emerges through prolonged exposure to literature, through listening, imitation, revision, cultural immersion, and eventually through thinking directly in the new language rather than translating into it.

The bilingual poet therefore occupies a unique position. They are not simply a learner. They inhabit two literary traditions, two systems of metaphor, two rhythms of thought, and often two emotional landscapes. Their challenge is not to erase one in favor of the other but to transform interference into artistic choice.

The following chapters examine how this transformation occurs from linguistic, literary, psychological, and computational perspectives.

Chapter I

Why Fluency Is Not Enough

Advanced learners frequently assume that literary writing is simply an extension of general language proficiency.

Experience suggests otherwise.

Many writers who speak a second language flawlessly continue to produce poetry that native readers immediately identify as foreign.

Curiously, readers often struggle to explain why.

The poems contain no obvious grammatical errors.

The vocabulary is sophisticated.

The syntax is correct.

Yet something feels translated.

This phenomenon cannot be explained by grammar alone.

Instead, it emerges from subtle interactions among collocations, conceptual metaphors, discourse organization, rhythm, register, sound symbolism, and literary convention.

The problem is therefore qualitative rather than quantitative.

It is not simply that the writer knows too few words.

It is that the writer’s aesthetic decisions continue to be organized by another language.

Chapter II

The Bilingual Poetic Voice

One of the greatest misconceptions surrounding bilingual literature is that writing in a second language inevitably produces diminished artistic expression.

History suggests precisely the opposite.

Some of the twentieth century’s most remarkable writers deliberately adopted languages acquired later in life.

Joseph Brodsky continued writing after exile in both Russian and English.

Samuel Beckett chose to write much of his mature work in French before translating himself into English.

Agota Kristof produced her celebrated prose in French, despite describing herself as linguistically displaced.

Vladimir Nabokov reinvented himself as an English-language novelist after beginning his literary career in Russian.

Each illustrates a different possibility.

None became “native.”

Instead, each developed an additional literary identity.

Their work demonstrates that bilingual creativity does not emerge from eliminating difference.

It emerges from mastering it.

The bilingual voice is therefore not an imperfect imitation of native speech.

It constitutes a legitimate literary mode with its own possibilities.

Chapter III

Thinking Through Metaphor

Modern cognitive linguistics has fundamentally altered our understanding of metaphor.

George Lakoff and Mark Johnson demonstrated that metaphor is not simply decorative language.

It constitutes one of the primary mechanisms through which human beings organize abstract thought.

We do not merely describe experience metaphorically.

We conceptualize experience metaphorically.

Languages frequently share broad conceptual patterns.

Time becomes space.

Knowledge becomes light.

Affection becomes warmth.

Yet every language also develops unique metaphorical networks.

The bilingual poet therefore encounters an unexpected challenge.

Translation transfers meaning.

Poetry transfers conceptual worlds.

Learning another poetic language requires constructing new metaphorical habits rather than merely expanding vocabulary.

Chapter IV

Prosody and the Ear

Every language possesses its own acoustic architecture.

English organizes rhythm primarily through stress.

Italian privileges syllabic continuity and vocalic flow.

Spanish occupies an intermediate position, combining syllabic regularity with distinctive stress patterns.

These differences extend beyond formal meter.

They influence breathing.

Sentence length.

Pause placement.

Expectation.

Silence.

The bilingual poet often continues hearing verse through the rhythmic habits of the first language long after vocabulary has become native-like.

Consequently, poems may remain recognizably foreign despite grammatical perfection.

Learning literary prosody therefore means acquiring a second ear.

Chapter V

When Computers Hear What Humans Feel

Recent developments in computational linguistics have produced an unexpected ally for literary criticism.

Stylometric analysis can frequently distinguish native from non-native writing even when expert human readers cannot explicitly identify grammatical errors.

Algorithms examine distributions rather than correctness.

Function words.

Collocations.

Sentence length.

Lexical diversity.

Dependency structures.

Information density.

Punctuation.

Rhythmic regularities.

The remarkable implication is that style contains measurable traces of linguistic history invisible to conscious inspection.

What readers describe intuitively as “sounding translated” increasingly proves computationally detectable.

Stylometry therefore offers empirical confirmation of phenomena long recognized by experienced editors and translators.

Chapter VI

Memory, Emotion, and the Multilingual Mind

Psychological research increasingly suggests that different languages activate partially different autobiographical networks.

Many bilingual speakers report that childhood memories emerge more vividly in one language.

Professional identity in another.

Romantic experience in a third.

Emotional intensity frequently varies according to linguistic context.

Some emotions become easier to articulate.

Others become strangely inaccessible.

Poetry inevitably reflects these psychological dynamics.

The bilingual poet is therefore not merely changing languages.

They are often accessing different versions of themselves.

Writing becomes an act of psychological translation as much as linguistic creation.

Chapter VII

From Translation to Creation

The ultimate transformation occurs when translation gradually disappears.

Not because the writer forgets the first language.

Rather because artistic intuition begins operating directly within the second.

Images arise spontaneously.

Collocations appear without conscious selection.

Rhythm becomes instinctive.

Revision shifts from correcting grammar toward refining aesthetics.

This moment cannot be scheduled.

It arrives differently for every writer.

Yet nearly every successful bilingual author describes some version of it.

The second language ceases to function as a code.

It becomes imagination itself.

Conclusion

The journey from language learner to bilingual poet cannot be measured by vocabulary size, grammatical accuracy, or examination scores.

Its true measure is aesthetic independence.

The writer eventually stops asking:

“How do I say this in Spanish?”

Instead, another question emerges:

“How would Spanish itself imagine this experience?”

That subtle change marks the birth of a second poetic consciousness.

The goal is never to erase the first language.

The goal is to allow two literary minds to coexist within a single writer.

Interference gradually becomes influence.

Influence becomes dialogue.

Dialogue becomes creation.

The bilingual poet ultimately discovers that writing in another language is not an escape from one’s original voice.

It is the discovery that the human imagination is capable of growing more than one.

Concluding Reflections: AI as a Teacher, Not an Oracle

If there is one lesson I hope readers will take from this series, it is that learning a language—and especially learning to write literature in one—is not a matter of accumulating vocabulary or memorizing grammatical rules. Those are necessary foundations, but they are only the beginning. The real challenge lies in acquiring an aesthetic intuition: learning to hear, feel, and eventually think in another language.

That realization led me to an equally unexpected discovery about artificial intelligence.

Much of the public conversation surrounding AI focuses on speed. People ask it for instant answers, quick summaries, recipes, itineraries, or the digital equivalent of a Magic 8 Ball. Used that way, it can certainly be convenient. But convenience is perhaps the least interesting thing about it.

Its real educational value emerges through sustained conversation.

Throughout the preparation of these essays, I did not simply ask questions and receive answers. I returned to the same ideas repeatedly. I refined prompts, challenged conclusions, changed perspectives, requested comparisons, and explored unexpected connections between linguistics, poetry, psychology, translation studies, and literary criticism. Each exchange became another draft of a much larger conversation.

In that sense, AI became less like a search engine and more like a patient interlocutor: one willing to explain, reconsider, elaborate, and follow an idea wherever genuine curiosity happened to lead.

This is not a substitute for reading books, studying languages, consulting experts, or engaging with the literary tradition itself. On the contrary, it has repeatedly sent me back to those sources with better questions than I had before. That, to me, is one of the hallmarks of good teaching. A good teacher does not merely provide answers; a good teacher helps us formulate deeper questions.

For polyglots, writers, translators, and lifelong learners, this may be the most promising role AI can play. It allows us to think aloud, to test ideas without embarrassment, to revisit difficult concepts as many times as necessary, and to connect disciplines that are too often studied in isolation. Used thoughtfully, it becomes a partner in intellectual exploration rather than a machine for producing shortcuts.

The essays in this series grew out of precisely that process. They were not generated in a single exchange. They emerged through dialogue—through questioning, revising, clarifying, disagreeing, and refining. If they contain insights worth sharing, they are a testament not only to what AI can do, but also to what curious learners can accomplish when they treat AI as a medium for inquiry rather than a source of instant certainty.

Perhaps that is the most valuable lesson of all.

Artificial intelligence does not diminish the role of the human learner. It magnifies it. The quality of what we discover depends less on the sophistication of the tool than on the depth of the questions we are willing to ask, the patience with which we pursue them, and the willingness to revise our own thinking along the way.

Ultimately, language learning, literary creation, and intellectual growth remain profoundly human endeavors. AI does not replace that journey. At its best, it accompanies it—encouraging us to read more carefully, think more deeply, and remain curious long after we have become fluent.

And finally, a personal comment to those readers who may be AI skeptics:

The quality of an AI interaction is often strongly shaped by the quality of the inquiry behind it.

A shallow prompt can certainly produce a shallow answer. If someone approaches AI as a vending machine—”give me a list,” “summarize this,” “write this for me”—they may get something useful, but they are often leaving much of the potential unexplored. The most interesting uses tend to involve iteration: asking why, challenging assumptions, requesting alternatives, examining trade-offs, comparing perspectives, and refining an idea over time.

In that sense, using AI well resembles several older intellectual practices:

• The Socratic dialogue, where the purpose is not merely to obtain information but to sharpen thinking through questioning.
• The workshop model in writing, where drafts improve through feedback and revision.
• The research seminar, where the quality of the discussion depends heavily on the quality of the questions being pursued.
• The relationship between a scholar and a good editor, where the editor does not create the author’s ideas but helps reveal and strengthen them.

There is also an important nuance: a weak question does not necessarily mean an unintelligent person. Sometimes people do not yet know how to formulate the deeper question because they are still at the stage of discovering the problem. One of the most valuable roles of an AI tutor can be helping someone move from:

“What is the answer?”

toward:

“What is the real question I should be asking?”

That shift is often where learning begins.

For example, as regards questions about writing multilingual poetry. The initial question might be:

“How do I avoid mistakes when writing poetry in Italian?”

A useful answer could list common errors. But the deeper chain of questions becomes:

“Why do these errors occur?”
“What does my first language impose on my second language?”
“How does metaphor change across cultures?”
“What does it mean to have a poetic voice in another language?”
“Can a person develop a second aesthetic intuition?”

Those are fundamentally different levels of inquiry.

At the same time, there is a reason some people are skeptical of AI, and some of that skepticism is productive. AI can produce confident-sounding but incorrect information. It can encourage passivity if used only for outsourcing thought. It can also tempt users to accept the first plausible answer rather than interrogate it. The difference lies partly in the user’s approach: curiosity, verification, and critical engagement remain essential.

Perhaps the most interesting educational possibility is that AI can become a metacognitive tool—a tool for thinking about thinking. A learner can ask not only:

“Teach me French poetry.”

but:

“What assumptions am I making about French poetry because of my English-language background?”

or:

“Why does this translation feel correct but still sound unnatural?”

Those questions reveal the hidden structures beneath knowledge.

A useful analogy might be this: a piano does not make someone a musician. But a good piano allows a committed musician to explore possibilities that would otherwise be difficult to reach. Likewise, AI does not replace intelligence, curiosity, taste, or judgment. It gives those qualities a new environment in which to operate.

The most exciting users are not necessarily those who ask the most questions. They are often the ones who learn how to ask better questions. And that, in itself, is a skill worth cultivating.

— Adam Donaldson Powell