It was my privilege on Monday morning, immediately upon our return from spring break, to have breakfast with our Cum Laude Society members of the Class of 2025 in the Olson Presentation Room, and in particular to enjoy reflective and hopeful speeches by Mac Froedge, Andrew Haas, Lilly Loeb, Alex McCarter, and Norah Wright. This year as every year, the Cum Laude gathering reminded me how little time remains for our seniors on campus—only 18 days to go (not that any of them are counting, of course)—and how proud I am of the people they have become and are still becoming.
To stand at a podium and speak with heart and mind as these five students did on Monday is a time-honored manifestation of an excellent education. The value we assign to public speaking at MICDS reveals itself in large ways and small—in the 108-year longevity of our Prize Speaking competition, and in the everyday ease of our students in all three divisions in addressing their peers from floor and stage alike. Respecting the latter, I have been particularly struck this year by our Upper School students’ liberation from their cell phones at the Brauer Auditorium podium and the consequent rise in “eyes up” delivery and clear vocal projection it has effectuated.
Occasionally I am asked whether MICDS students today can distinguish between more and less formal spoken diction given the ongoing decline in adherence to traditional grammar, the erosion of specificity, and the rise of “filler words” in spoken English (“like,” “um,” “you know”) over the last several decades. Curious about the answer, I decided to conduct an experiment with our students in grades 7-12, and I am grateful to the several hundred of them who chose to participate.
I began by writing a brief bit of first-person prose:
We arrived at an open meadow after hiking for about an hour. Wildflowers abounded — I don’t think I had ever encountered so many wildflowers in one place — purple and yellow and blue, wafting in a breeze that carried their earthy scent. The tall, tufted meadow grasses swayed in it too, and I experienced a feeling of utter serenity in that moment.
I then wrote a second version of the text, changing some of the language from less common to more common words and phrases but otherwise leaving it unaltered:
I remember that we came to an open meadow after hiking for about an hour. There were so many wildflowers — I don’t think I’ve ever seen so many wildflowers in one place — purple and yellow and blue, and I remember there was a breeze, and I smelled the wildflowers in it and watched the tall grasses swaying in it, and a feeling of complete peace and joy came over me in that moment.
Finally, I wrote a third version that retained the text of the second but added filler words throughout:
I remember we came to like an open meadow after hiking for about, like, an hour, and there were so many wildflowers — I don’t think I’ve ever seen so many wildflowers in, like, one place — purple and yellow and kind of blue, and I remember there was like a breeze, and I smelled the wildflowers in it and like watched the tall grasses swaying in it, and a feeling of like complete peace and joy came over me in that moment.
With these three variants in hand, I used ElevenLabs AI technology to record each of them separately with a female and a male voice. (Here they are if you’d like to hear them yourself: female version one; female version two; female version three; male version one; male version two; and male version three.) I then divided students randomly into six groups and, without giving them any of the foregoing information, simply asked them to listen to the single audio clip that I had assigned to them and rate the speaker as “extremely intelligent,” “somewhat intelligent,” “neither intelligent nor unintelligent,” “somewhat unintelligent,” or “extremely unintelligent.” After converting their responses to a 5-point scale, not surprisingly—and perhaps encouragingly for people who worry about the “dumbing down” of spoken English—I found that our students, on average, associated more descriptive vocabulary with intelligence, especially in the absence of filler words. In fact, they assigned almost twice as great an “intelligence penalty” to the use of unnecessary “likes” than they did to the use of less descriptive word choices.
Digging into our students’ responses further, I was fascinated to discover a slight correlation between gender affinity and perceived intelligence. Male students who listened to the male AI and female students who listened to the female AI were more likely to hear intelligence in those voices than female students listening to the male AI or male students listening to the female AI. The difference was not significant, but it was noticeable in the data; and while this little experiment on my part would hardly bear up under scientific scrutiny, at a minimum I am comfortable observing in a general way that we should always be mindful of the unconscious biases we bring to our interactions with other people.
It would appear not only that today’s MICDS students can distinguish between more and less traditional diction, but that their attribution of rhetorical value to word choices and phrasing is directionally consistent with that of prior generations. Mark Twain would have expected nothing less. “When a great orator makes a great speech,” he wrote to Helen Keller in 1903, “you are listening to ten centuries and ten thousand men—but we call it his speech, and really some exceedingly small portion of it is his.” It will take a lot more than TV and TikTok to undo such a profound oratorical inheritance. Which is, like, a good thing, you know?
Always reason, always compassion, always courage. My best wishes to you and your families for a happy and restful weekend.
Jay Rainey
Head of School
This week’s addition to the “Refrains for Rams” playlist: The Kids are Alright by The Who (Apple Music / Spotify)