The Thing You’ve Been Saying Your Whole Life

Julie Sedivy reviews Michael Erard’s “Bye Bye I Love You: The Story of Our First and Last Words.”

By Julie SedivySeptember 6, 2025

• Cultural Studies

• Science & Technology

I WAS FOUR years old when I first witnessed the green shoots of a baby’s emerging language, that of my younger brother. Since I belonged to a large extended family and then had my own children in my twenties, the first words of toddlers were a frequent milestone, heavily discussed and analyzed by my word-loving clan (and by my linguist friends and colleagues). But it wasn’t until I was in my fifties that I witnessed someone’s last words for the first time—by sad happenstance, those of the same brother whose first words I had been privy to, as he lay dying of pancreatic cancer.

This asymmetry in my own experiences of first and last words mirrors a broader cultural reality. As Michael Erard notes in his wise and gracefully written new book, Bye Bye I Love You: The Story of Our First and Last Words, first words have invited intense scrutiny and scientific attention, whereas last words live in the shadows. In a society that prizes verbal skill, we’re conditioned to see the beginnings of language as the child’s entry into communal society, perhaps a hint at their future talents and inclinations. The early utterances of children have been harvested as data to decide among competing theories of how language originated in humans and how articulate minds work. The words of the dying, on the other hand, have held little interest outside the private sphere, except when compiled as the final aphorisms of the famous. What is there to say about the disintegration of language?

But Erard looks beyond this asymmetry to find meaningful parallels between first words and last and, in doing so, highlights the human talent for connection that undergirds language from its beginnings, through its apex, and into its decline. First and last words occur at stages of life when linguistic ability is most fragile—when language’s forms have not yet consolidated or when they’ve begun to break apart, whether due to dementia, aphasia, or simply the “neurochemical commotion” that occurs as the body is shutting down. It’s when the language code is least reliable that it becomes most visible how intimate an act it is to communicate.

As Erard points out, communication is not just about mastering the forms of language; it’s also a collaboration. We coordinate attention with each other. We guess at the contents of each other’s minds, at times based on the slightest of linguistic signs but drawing from a deep well of knowledge; we propose meanings and accept or reject them. “Doo!” my daughter exclaimed, while stepping her tiny foot into my shoe. “Yes, that’s my shoe,” I said, mentally noting that her pronunciation aligned perfectly with the sound distortions known to be typical of toddlers. “Doo!” she repeated with fervor, which I took as affirmation of my understanding and a celebration of her linguistic success. This is how I remember the first word she uttered.

But was it really? Erard might ask. Perhaps her first word was the general-purpose syllable Da! spoken while pointing at anything she wanted me to notice with her. Or perhaps a first “word” is any agreed-upon sign that need not be verbal at all. In his book, he relates how a 12-month-old girl pointed to a rosebush in the garden and snuffled, an imitation of a gesture she had seen her mother make while handling a flower. Is that a word? wondered her mother. Is she naming the flower? The snuffle certainly seemed to be performing the same function as a name. Erard is spot-on when he remarks that “a baby’s first word isn’t a ribbon-cutting of their language project but the tip of an iceberg’s worth of linguistic knowledge that they’ve already acquired.”

A similar intimacy envelops language at the end of life, when the words that a dying person can produce may be the tip of an iceberg’s worth of what they want to convey. Here, too, the interlocutor needs to lean in close, ready to propose and respond to interpretations not just of words but also of any act that seems to carry intent and meaning: a touch, a sigh, the particular quality of a moan, or even just a fluttering of the eyelids. “First and last words,” writes Erard, “are often taken to be singular expressions of the individual person. However, in linguistic terms, responsibility for them is shared. We negotiate them; we make them together. In that sense, your first and last words are never yours and yours alone.”

In fact, current research shows that language at all stages of life is collaborative in its essence. Even with an ample vocabulary and an agile syntax, people rely on their shared histories and their calculations of each other’s abilities or intentions to untangle the ambiguities that riddle language and to read beyond what is said to the heart of what is meant. Listeners emit subtle signals to indicate their level of understanding, and attuned speakers adapt accordingly, adding more elaboration when needed, moving the conversation forward when alignment has been achieved. Without these interactive skills, language would need to be robotically explicit, all meaning sitting on the surface; there would be no innuendo or sarcasm, no comedy that relies on setting up unspoken expectations and then knocking them down. (I’d like to die peacefully in my sleep like my uncle, goes the joke. Not screaming in terror like his passengers.)

Erard gestures at this interactive turn in contemporary linguistic research, recognizing that scientific interest has shifted over time from a narrow preoccupation with the structure of the linguistic code to an appreciation of its cooperative, embodied practice among social creatures. His own discussion, though, is firmly fixed on those early and late stages of life when the language code is especially brittle. Incidentally, this is also where language remains the most recognizably human. Current language models that power artificial intelligence are remarkably good at replicating aspects of the language code. But no one has built a chatbot with the social depth or subtlety to respond to a child’s first words or to grasp the import of a person’s last.

To be sure, there is much uncertainty in these interactions, when the words themselves give us so little, leaving plenty of room for misunderstanding. To fill in the many gaps, we drag in not only our knowledge of each other but also a cartful of cultural expectations. And our society’s norms and values determine whether we burden these utterances with significance in the first place. One of the most rewarding aspects of the book is the way in which Erard draws attention to cultural variations that shape these encounters.

An obsession with children’s early words turns out to be anomalous even within Western culture across time, a product of the recent intrusion of scientific expertise into child-rearing practices. In many societies, adults take little notice of children’s early words; they are simply absorbed into the flow of daily life. Of his own preoccupation with his children’s speech, Erard writes, “We’re the sort of American parents enculturated to believe that our offspring are, among other things, cognitive projects. In the parenting literature we encountered, no baby was perfect enough. They were always improvable.”

Cultural expectations set the stage at the end of life as well. In some religious communities, last words are ritualistically prescribed, a good death being one in which a specific prayer or the name of a god are the final words to leave one’s lips. A deathbed prayer may have been seared into the mind by endless repetition, beginning in childhood. In contemporary Western society, we expect not just an unscripted burst of spontaneity but even something profound: a secret divulged, a final summation of one’s life, a forgiveness finally granted.

These expectations can collide with the physical realities of dying, cautions Erard: “Raised to believe that dying people normally exit with a quip, a teaching, or a personal moment, you’re left with no framework that accounts for the fragmented, the nonsensical, the delirious, or the silent. You expect to raise a monument to the mortal self; what you trip over is rubble.” A good portion of the text is devoted to documenting the clinical facts of death and the ways in which they undermine the “glowing linguistic powers” that we wish for at the deathbed. For example, among a group of German palliative patients who were sedated, as is common, only half were able to communicate at all, even to respond to simple questions about their pain level or physical comfort.

Erard’s prose is tinged with sensitivity throughout, but in these particular sections, he writes from an intellectual remove that may be jarring for some readers, especially if they come to the book with the fresh wounds of a loss. But perhaps this, too, reflects our cultural norms; we are comfortable with an analytical dissection of the words of toddlers but tend to recoil from explicit accounts of the process of dying.

It is all the more vital, argues Erard, to question our cultural assumptions, given that it has gotten more likely, not less, that people will die under conditions of greatly diminished linguistic powers. In the past, many people died of infectious disease, often at a young age and without the medical resources to keep people alive in a prolonged state of cognitive incapacity. Today’s dying tend to be older, often suffering from dementia or the effects of a stroke, or from multiple illnesses that gradually close down their functioning. And sedation to relieve discomfort or agitation has become commonplace.

A striking omission from this discussion is the growing practice of medically assisted dying, which introduces an entirely new dynamic into end-of-life interactions. In these situations, one knows the precise moment of death, can anticipate exactly which words will be the last. Alertness and a degree of linguistic competence are often required. On one hand, the dying and their loved ones have a great deal of control over the conditions of death, allowing for lucid, meaningful last exchanges; on the other hand, the weight of cultural expectations in these circumstances could well be crushing, a source of great pain in the aftermath if the final words fail to live up to the moment.

Whatever the conditions of death, I’m inclined to embrace Erard’s friend Louis’s suggestion that “your last words shouldn’t be the first time you share a truth or tell a secret, they should be the last time you say the thing you’ve been saying your whole life.”

This piece of advice brought to mind the last words that my dying brother uttered, how they were the echo and distillation of so many of his previous words. In the months after his diagnosis, I spent many hours on the phone with him. In these rambling conversations, he combed through his life’s history, wondered about his legacy, tested out his beliefs about death and whether anything came after. He always came back to the theme of love: amassing evidence for it throughout his life, asserting that, whatever he did or did not believe, he believed in love.

In his final days, he could manage little in the way of talking. He slipped in and out of delirium and consciousness. When it seemed clear that the end was near, his closest friend, whom he had known since junior high, began reciting every shared memory he could think of. When he got to the end, he started all over.

And the one word my brother mustered in response to all this was: “Love.” He repeated it again and again. “Love. Love. Love.” His friend sent it back to him: “Love. Yeah, love.” This back-and-forth volley of a syllable between them was the very embodiment of the word itself, a word that contained everything of importance to my brother from beginning to end.