One of the pleasures of returning to Oxford after a few decades spent far away is that I get to support younger versions of myself, helping them to gain experiences that I wish that I had myself as an undergraduate. A case in point is the undergraduate research programme that I ran this summer as a bit of an experiment, together with some of my Maryland PhD students, plus some crucial help from colleagues in Oxford. 

At the start of May, I was honestly unsure whether any students would be interested in taking part. At the start of June I still had only a vague idea of how the programme would work. Now, at the start of August, we can look back on an exhilarating 8 weeks, where part of the fun came from the fact that we were “building the airplane while flying it”. 

This is a brief summary of why we did this, how it worked, what were the outcomes, and the surprises, and some take-aways for the future.

Motivations

My experience as an undergraduate in Oxford was intellectually stimulating. But it was also isolating. And I didn’t have a clear sense of purpose, beyond the excessive focus on getting good grades. (I didn’t know any better at the time. It reflected the values around me.) This was a key reason why my plans pivoted so quickly when I landed in Rochester, NY in 1990. The sense of shared intellectual purpose, across people from different backgrounds, was infectious. I quickly decided that I wanted more of this, and that to get it I would need to stay in the US for a while longer. That turned into a lot longer.

In my time in the US, I saw again and again how much students can benefit from early research experiences. I saw this in students in MIT’s UROP program. At the University of Delaware, a couple of the undergraduates who worked with our research team won Rhodes Scholarships. At the University of Maryland I saw countless undergraduates have fabulous research experiences. Due to my focus at UMD on graduate training, I was more of an observer than a contributor to undergraduate research. But I worked with a long line of fabulous post-bacs, and so many of my graduate students benefited from undergraduate research experiences themselves.

In the US, undergraduate research is a fairly normal part of university culture. Students are less fixated upon grades — there is no binary cut off between the two degree grades that almost all students get in the UK (a “first” or a “2:1”, on everything) — and students are more interested in building a portfolio of different experiences. 

I wish I had similar experiences myself as an undergraduate. For years I have been saying that, if I found myself in the UK again, I would like to do something about this. I may have mentioned it in job interviews, too. And now here I am! 

To my surprise, in doing research on the 20th century history of Oxford, I learned that earlier undergraduates may have been closer to the intellectual coal face than my generation and beyond, in part due to the relatively recent expansion of graduate education. Nowadays, graduate students and postdocs outnumber undergraduates in Oxford. Understandably, they are closer to the research action. But through the mid 20th century, undergraduates were a more dominant group, and many academics did not have graduate degrees. For this and other reasons, this probably made it easier for undergraduates to be closer to the action. 

To my surprise, while researching material for a memorial talk for Janet Dean Fodor, one of my unofficial mentors, and also a Somervillian (1961-64), I learned that she had done substantial undergraduate research herself in Oxford. In fact, by far her most cited work ever is a paper resulting from that undergraduate work, joint with Michael Argyle. (Almost no psycholinguists have heard of the paper, because it appears in a different field, social psychology, and under her unmarried name.)

As an additional motivation, I wanted to understand more about the experiences of current Oxford undergraduates. Maybe my experience from decades ago was just hopelessly out of line with current students’ experiences. I needed to find out.

Also, I was interested in finding specific ways to help build bridges between my students in Oxford and Maryland. Some of my Maryland PhD students hoped to spend part of the summer in Oxford. I wanted to find ways for them to be integrated, to contribute, and to also benefit their own development.

Troubleshooting during an early session on data wrangling

Inspirations and Support

Our summer research programme did not occur in a vacuum. It took inspiration from a number of other summer activities that have left an impression on me. 

For the past 3 summers I have been involved in various ways in the Language Science Station summer program at the Planet Word Museum in Washington DC. Students in that program get to carry out research studies live on the museum floor. The focus of the program is mostly on the communication aspect. 

Language Science Station team, summer 2022. Undergraduates from Washington DC area universities take an intensive summer course in running research studies and engaging with the public, live on the museum floor at Planet Word museum.

I have also taken some inspiration from the long-standing CSLP Summer Workshops at Johns Hopkins University, which for 30 years has been bringing together teams of faculty and advanced students to work intensively together on some topic in speech and language engineering.

It also helps that, in my research area, recent technological developments have made it dramatically easier to gather human data online. My own team has developed tools that allow us to collect and analyze speech data in a matter of days that just a few years ago would have cost us months of work. This makes it possible to get more done in a short summer session.

Finally, it wouldn’t have been possible to get this summer’s programme off the ground so quickly without guidance and encouragement from Claire Cockcroft, director of the Somerville Skills Hub. We were introduced soon after I arrived in Oxford a few months ago. Claire helped to shape my vague ideas and helped me plug into student needs. We also could not have gone ahead without support from colleagues in reaching interested students, and from some colleges who came through with support for participating students.

The taste of victory – getting the first experiment to work (Week 8)

What Happened

In short: we worked as a team, taking a research project from the very earliest stages to having an end in sight, over the course of 8 weeks.

Our project focuses on the cognitive processes involved in turning ideas into spoken phrases and sentences. Our main tool involves tracking fine timing differences in speech as people describe simple scenes. We collect the data from participants online, and we analyze the speech using a semi-automated pipeline that uses AI tools to transcribe speech and to measure the durations of sounds and words. Our specific focus was some recent interesting-but-surprising findings by Shota Momma and others (myself included!) that suggest an especially tight relation between the planning of verbs and their objects in speech planning.

We started the programme with only a vague idea of “this is an issue that we’d like to understand better”. 8 weeks later, we had 3 parallel experiments ready to go, on English, Mandarin, and Czech. The undergraduate students were involved in every aspect of planning and implementing the project. 

We were 8 Oxford undergraduates, 2 Oxford MPhil students, 2 UMD PhD students (Rosa Lee, Allison Dods) who were with us for most of the programme, plus a couple of others who helped out at key moments, either in person or from afar (due to visa issues). So, at full strength we were more than a dozen people.

I was physically present for the full 8 weeks of the programme. I planned to be in Oxford for the summer anyway. Everybody else had made plans at shorter notice, so between other travel commitments (jury duty!), exams, illness, etc. most participants were with us in person for 6-7 weeks in total.

We quickly found a weekly and daily rhythm that worked well for about the first 5 weeks of the programme. After that point we had clear enough objectives that we were able to adjust to needs on a day-to-day basis. 

On a typical day we would have a structured morning session, initially focusing more on the science. Those sessions were generally led by some combination of myself, Rosa, or Allison. Sometimes these were run more like a seminar, other times these were run more like a workshop focusing on a specific problem. 

In a typical afternoon, students would work together on developing specific skills, such as how to code up an online experiment, or how to automatically analyze speech recordings. Sometimes we patched in Utku Turk from the US to help with coding or quantitative training.

On a weekly basis, we laid out a small set of goals for the week, which we would discuss with participants each Monday morning. These had to be figured out on the fly, so a fun feature of the programme became our weekly Sunday evening “pub & planning” meeting among me and the grad students. Each week we sampled a different one of Oxford’s many pub gardens. And we got a lot done!

Sunday evening “Pub & Planning” meeting at The Perch in Binsey. We were almost the only ones there that night, as England was in the final of the Euro 2024 soccer tournament (don’t ask).

Here’s a rough outline of what we did week by week.

• Week 1: Calibration and familiarization. We needed to get to know each other, and figure out what the students know. Students also worked as a group to implement a simple picture-word interference study in PCIbex. And we had a guest presentation, by Mal Shah (Oxford alum and UMD PhD student), on the semantics of verb-argument relations. 
• Week 2: Refining our understanding of the psycholinguistic and analytical linguistics background; completing implementation of initial experiments
• Week 3: Assessing in detail the specific questions that we could pursue. Some approaches that seemed attractive at first had to be jettisoned. Learning the data analysis pipeline
• Week 4: Choosing specific questions, preparing a group presentation for the end of the week; refining data analysis skills.
• Week 5: Turning experiment goals into specific experimental logic; more analysis skills development; Q&A session about graduate study.
• Week 6: Turning experimental logic into detailed plans for 3 parallel experiments, in English, Mandarin, & Czech.
• Week 7: Getting deep into the weeds on experiment parameters, materials, pre-pilot testing.
• Week 8: Putting final touches on pilot experiments, initial data collection and evaluation.

We did not know at the start of the programme where we would end up by the end. For example, we never would have predicted that 2 of our 3 experiments involved Czech adjective-noun production and Mandarin classifier-noun production. (We had speakers of both languages in the group whose native language skills we could tap into.) But this choice made perfect sense as our understanding evolved.  

The undergraduate students were involved in the decision making process every step of the way. They read challenging papers in semantics (by Kratzer) and psycholinguistics (by Momma), among many others. They picked apart experimental data from a relatively new experimental paradigm that we wanted to use. They wrestled with questions of how to get automated speech analysis to work with “smaller” languages such as Czech. They puzzled over how differences in the prosodic constraints of Mandarin and Czech might lead experiment participants to respond differently to pressure to speak quickly.

This was definitely not the kind of situation in which undergraduates are assigned to be low-level assistants in already-designed projects.

The goal of giving a group presentation to faculty and graduate students in a lab meeting at the end of Week 4 proved to be a valuable milestone. It was the point where students really took ownership of the project, as it pushed them to be really clear about their understanding, and to advocate for the choices that we had made collectively.

Working at The Hub at Kellogg College on preparing for the group presentation (Week 4)

Outcomes

All indications so far are that the programme achieved what we wanted, and more. 

We have 3 closely related experiments ready to go, in 3 different languages. At the time of writing, we are in the middle of pilot data collection. We will continue to complete the studies, which we hope will lead to presentations and maybe publications that the students will be full partners in.  

The undergraduate and masters students gained many new skills, and they learned how to put them together in the service of addressing difficult research questions. We were learning together how to use new experimental approaches and new tools, and the students were full partners in this process.

In the case of technical skills in coding and data analysis, the students started with varying levels of experience and confidence. They progressed further than we expected, they became more self reliant than we had expected, and they enjoyed it more than we expected. 

Students gained a better appreciation of the many different details that make a difference when carrying out psycholinguistic research. This is not so much about experimental hygiene, and more about carefully working through linking hypotheses, understanding how underlying cognitive processes lead to observable distributions of data, thinking through how participants would approach a task, and more. At first, it took some work to convince the students that these details really matters. But it eventually clicked. Whereas nobody was surprised that things like coding and quantitative skills are useful, most students were surprised to learn about the many other details that matter a lot.

One outcome that proved to be more important than expected was teamwork. I knew that my own undergraduate experience in Oxford had been isolating. But maybe this was because I was in a small, obscure field (mostly studying medieval German literature). It seems that my experience was not an outlier. The current students get a lot of individual attention from faculty. They are very well supported. But they were not accustomed to working together. They took to this really well. And they really enjoyed it. For some, it created a sense of belonging to an academic community that they had not felt before.

All of the students took part because they were curious about doing research in the future. All of them came away with a clearer understanding of what this involves. And they were overall encouraged by what they saw. One commented that they had heard various horror stories about research, but that it was much more fun than they had expected.

For the PhD student mentors, the experience was also a positive one. They got to develop ideas that they were interested in, on an accelerated timeline, working with a bunch of other smart people. So it advanced their research, and it provided valuable experience in designing and leading programs. They were able to make many new connections, and to integrate themselves well into a different research community. And it did not hurt that they got to do so in the rather nice setting of Oxford in the summertime. 

So far, these are only impressionistic reports. Soon we will have more evidence, from a pre-/post-survey of participants carried out at the start and end of the programme, with some guidance from Oxford’s Centre for Teaching and Learning. This is the gold standard for measuring whether short interventions like this really do make a difference. But based on participant reports so far, the evidence is already rather clear.

So many octopi – an intense discussion about speech planning mechanisms

Surprises

The biggest surprise was that we managed to pull it off at all, at such short notice.

I was surprised at the level of hunger from students for this kind of experience. At the start of May I was worried that nobody would want to take part. Just one week later I was more worried that we had more interested students than we could handle. And I didn’t even need to work that hard to recruit. (To be clear, a couple of colleagues gave me some valuable leads on students who might be interested.) 

I was not surprised that the students proved to be rather talented. That goes with the territory. But I was surprised, and I think the students were surprised, by what it took to unleash those talents on research. I think we were all struck by the gap between the skills they use in the rapid cycle of weekly essay, tutorials and classes, and what they need to really sink their teeth into a research problem.

We were all pleasantly surprised by how well the group functioned as a team. Given the range of backgrounds (in linguistics, psychology, and philosophy), ages (1st years to masters students), and personalities, we were unsure how well this would go. We were very explicit from the outset that team dynamics are important, and we provided some structure to create opportunities to collaborate. But it progressed better than we had expected.

I was pleasantly surprised how far the students progressed in just 8 weeks. They were not ready to dive in as researchers on Day 1. But 8 weeks later they were much more ready. Some students commented that the experience would make a big difference to how they would approach the rest of their degree. Another said they wish they had this opportunity much earlier, as it would have changed their approach to their studies.

Taking stock at The Jericho Tavern (Week 8)

Future

We framed this year’s summer programme as an “experiment”. We weren’t sure if there would be interest, and whether or how it could work. 

Those questions have now been answered. There is a clear hunger from students. The programme succeeded beyond what we could have hoped for. We have found ways to make it work well.

In principle, then, we could hope to offer more of this in the future, possibly scaling up to broader participation. This aligns well with university priorities for enriching student experiences in ways that go deeper, that are more interdisciplinary and collaborative, and that provide better preparation for career options after graduation.

There are a few things that could make a big difference in our ability to sustain and scale a program like this. 

1. Student support. Some of this year’s participants received support from their college to help them to stay in Oxford during the summer. Finding additional sources of student support would make a big difference.
2. Mentor support. I simply could not have pulled off this programme on my own. Doing it together with Rosa Lee and Allison Dods (on site), plus remote help from Utku Turk, was hugely important. It also provided valuable experience for the PhD mentors. The main costs, which I covered this year, were their travel to Oxford, and accommodation during their stay. (My own involvement as programme leader cost nothing.) Finding ways to support the co-leaders is essential.  
3. Space. We were lucky that the Clarendon Institute, the slightly worn and soon-to-be-former home of Oxford Linguistics, is a bit of a ghost town in the summer. So we were able to easily find space that we could use for the programme for 8 weeks, basically as squatters. It is unclear whether this will be possible after we move in 2025 into the beautiful new Schwarzman Centre for the Humanities. There are incentives for the university and colleges to monetize every inch of space in Oxford during the summer. 
4. Aligning values. Notwithstanding a sincere push to broaden student experiences, Oxford’s fixation on final exams and degree classifications risks undermining efforts like this one. Maybe a program like this doesn’t really count as educational because it is not directly preparing students for exams. Or maybe it gives participants an unfair advantage in the competition for prestigious degree grades. I was surprised how often issues like this came up when I talked with colleagues about plans for the programme. This is quite unlike how people think about undergraduate research in the US. 

I am extremely grateful to everybody who took a risk on devoting a big chunk of this summer to taking part in this “experiment”. It was a lot fun and very rewarding for me. I am looking forward to continuing the research that we got started. And I am equally looking forward to talking with anybody who is interested in getting involved in the future, or finding ways to support this kind of student experience.