The Importance of Card Messages in Expressing Condolences

Most people stall at the card. Flowers picked, address entered, then the box appears asking what the message should say. The order takes four minutes. The card takes twenty.

I went looking for research that would tell me what to put in that box, and what I found was uncomfortable enough that I threw out my first draft. The discomfort turned out to be the useful part.

I help run the US side of Lily's Florist alongside Andrew, from a small office in North Carolina. Joan alone has thirty years on the bench and eight on the phones, so we have opinions. Opinions are cheap. I wanted to know whether the thing we sell a hundred times a week does anything at all.

The short answer is that a condolence card does something real. What it does is narrower and stranger than what most people are attempting with it.

There are only three versions of this problem, and knowing which one you are in removes most of the paralysis. You knew them both. You know the griever and never met the person who died. Or you are signing for an office. I will get to all three.

In 2017 a French intensive care group took 242 relatives of patients who had died in an ICU, randomly assigned half to receive a personalized handwritten condolence letter from the medical team a month later, and measured their symptoms at six months.

The letter group did worse. Post-traumatic symptoms showed in 52 percent of them against 37 percent of those who got nothing. The authors wrote that the letter failed to alleviate grief and may have worsened depression and PTSD-related symptoms.

Be careful with that. A formal letter on institutional letterhead, from clinicians the family may have known for six days, arriving unprompted at week four, is a specific object. Your handwriting on a card to your sister is a different one. But this is the closest thing to a controlled test that exists, and it kills a comfortable assumption: that written condolence soothes by default, and that more of it is better.

So why does the card still matter? On the evidence it lowers nobody's symptom score, and nobody has measured the thing a card is actually for, which is being present from somewhere else. What the research carries is the other half: a large body of work on which sentences bereaved people report as wounding. The damage a bad card does is documented. The good a careful one does is not. That asymmetry is what the rest of this is about.

Two findings sit under this section. The first is about the brain of the person you are writing to. An early fMRI study of eight bereaved women found grief cues lit up the dorsal anterior cingulate cortex and the insula, regions other research ties to the distressing part of physical pain. Small study, and it lines up with what people report: grief is felt as something closer to injury than to sadness.

The second is about language. In 2007 a UCLA group showed that when people put a feeling into words, amygdala activity dropped and right ventrolateral prefrontal activity rose. Naming an emotion measurably quiets the alarm system. The effect is called affect labeling.

A quieter reason the format helps: a grieving person has very little capacity left for open-ended input, and a card demands none of it. Read at 2am, put in a drawer, reread in March. It asks nothing back.

In 2014 two Stanford researchers, Birgit Koopmann-Holm and Jeanne Tsai, went and bought sympathy cards. Not surveys about cards. The actual product. They analyzed 338 American and 376 German cards and counted what was inside.

American cards averaged 2.90 negative words. German cards averaged 7.30. Imagery of dying appeared in 2.53 percent of American cards and 16.26 percent of German ones. Four follow-up studies traced the gap to a cultural trait they call avoided negative affect: how strongly a culture believes it should not dwell in unpleasant feeling.

Reach for the pre-printed sentiment and you will produce something that steers the reader out of their own feeling. That is the American default, and it carries a cost the sender never sees.

Twenty bereaved parents were asked what people had actually said to them after their child died. Their answers are the most useful material here. The same study coded 170 Hallmark sympathy cards.

Across the card corpus, "love" appeared 104 times, "comfort" 84, "sympathy" 77, "heart" 74. That is the commercial vocabulary of American condolence. When the parents described what had actually reached them, they reached for a different set of words entirely.

What they named as unhelpful clustered into four types: religious explanation, grief comparison, unsolicited advice, recovery timelines. "It was God's will." "I know what you're going through, I lost my mother." Communication research calls that axis person-centeredness, and a 2012 meta-analysis found high person-centered messages reliably outperform low ones.

One situation raises the stakes. Kenneth Doka's term is disenfranchised grief: a loss the world has decided does not qualify. An ex-spouse. A partner nobody acknowledged. A pregnancy. A death by overdose. When society quietly declines to mourn a loss, a card stops being a courtesy and becomes evidence. For some people it is the only formal acknowledgment they will ever receive that something happened.

Nine grieving people in ten locate the feeling in the same place: the chest. A 2024 Swedish study of 116 bereaved people put that figure at 91 percent, and found 63 percent reporting a diminished sense of their own body in the first six months. Asked where a hug registered, 88 percent said the chest.

That is the practical case for paper over a screen, and it rests on where people say they physically feel things. The film theorist Laura Marks calls it haptic visuality: the eye working like a hand, grazing a surface rather than surveying it. Heavy stock, ink pressed into paper, weight in the hand, still on the windowsill in six weeks after the phone messages have scrolled away. The Victorians engineered this deliberately, with a hand-drawn black border that narrowed as mourning progressed, so the recipient knew what the envelope held before opening it.

I cannot give you a study with a decay curve on it, because as far as I can find nobody has published one. What exists is adjacent and real: a 2020 Oxford study found that feeling cut off from others drove distress after a bereavement, independent of the loss itself.

What I can give you is our own order data. It is the only part of this article I can see with my own eyes. Sympathy orders to a household collapse after about the second week. Almost everything arrives around the service. Then the house goes quiet, and the work starts: the paperwork, the accounts, the first weekend nobody visits.

Most people who send in week five apologize for it. They meant to do it at the time, the week got away, and by the time they remembered it felt too late. Late is the good version, and waiting costs nothing. If you want something still in the room after the cards come off the windowsill, this is the one place a plant beats cut flowers on the merits. A dish garden is alive in month three. Joan's line on it is better than mine.

A late send also removes the etiquette problem. Traditional practice says the bereaved should write thank-you notes weeks afterwards, turning a kindness into an obligation for the person least able to carry it. Cancel the obligation on the card, in six words.

One practical note. Sending to the house is a different product from sending to the service. Sympathy flowers for the home arrive in something the family does not have to find a vase for, at a moment when finding a vase is beyond them. Something quiet like Thinking Of You does it without announcing itself.