Curriculum 

Instead of removing or reducing an existing requirement, can’t we instead increase the number of units for an MIT degree, or free up space by REST, or add new requirement X to the majors?

We heard from some community members about the desire to find space within the existing undergraduate academic program for the new curricular elements proposed by TFUAP (such as Computing, PSM, or Team-Intensive) in order to avoid having to reduce the size of existing curricular elements. We examined many options for finding such space, and here provide a summary of the design constraints that we encountered and our rationale for our choice in the proposed solution.

The MIT undergraduate academic program, as stipulated by Rules and Regulations (section 2.84), is currently 32-34 twelve-unit subjects. This implies that for a four-year (8-term) degree program, in most terms a typical student takes four subjects (48 units), and in two of those terms a student takes five subjects (60 units).

One way to create space is to increase the expected number of subjects or units that students take each term. Since we often hear that students take more than four 12-unit subjects each term, we investigated whether students already take more than 48 units each term; if so, it would seem reasonable to use the actual number of units taken as the norm. We did not find evidence that students are taking much more than an average of 48 units per term, therefore adding more expected subjects per term would increase what we require of students. We obtained trimmed means (means trimmed by 5%) registration data from the Registrar, which indicated that students take an average of 50 units/term, with no discernible temporal trend over the past 20 years (except for a brief increase during COVID). 50 units/term comes out to ~33 subjects over 8 terms, right in line with Rules and Regulations. Thus, without requiring students to do more than they are already doing and have done for the past 20 years, there is no free space to be found by adding more expected subjects per term.

Another way to find space is to add another semester or year to the undergraduate program. This change is functionally equivalent to increasing the number of undergraduate students and would require additional dormitories, an additional semester or year of financial aid, additional classrooms, etc. We thus did not pursue that approach.

Increasing the number of units required for an MIT undergraduate degree, which is currently the GIR subjects plus 180-198 units, is equivalent to either increasing the number of subjects/term or increasing the number of terms, as discussed above, and thus, we did not pursue this approach either.

In summary, we did not attempt to find space by increasing the overall “footprint” of the MIT undergraduate program. Instead, we must look at how that footprint is currently distributed.

The entire undergraduate academic curriculum currently consists of 6 SME GIR subjects, 2 SME REST subjects, 1 Institute Lab, 8 HASS subjects, up to 12.5 subjects in the major, and 4 unrestricted electives. Adding those up comes to 33.5 subjects. 

One suggestion from some in the community is to reduce the number of HASS subjects and re-apportion them to the SME GIRs. As described in our report in the “Why Eight Subjects?” (p. 47) section, after extensive deliberation, we concluded that eight HASS subjects was appropriate.

Another area to look for space would be the majors. Several engineering departments communicated that they felt disadvantaged with the current major size, and the overlap rules for GIRs mean that it is not possible to create space by replacing REST/Institute Lab without negatively affecting majors. 

To understand the issues here, we must explain the Institute’s overlap rules. Above, we stated the Institute rule that courses may include up to “12 and one-half subjects and 150 units” (section 2.84.E.3), but many majors are in fact larger than 12.5 subjects. That is because Institute regulations allow for double-counting, or overlap, between the REST/Institute Lab/HASS and the majors, up to 36 units (3 subjects, section 2.84.E.3): 

“Departmental programs may specify or expect up to three subjects that are also used by students to satisfy the General Institute Requirements, with the understanding that the department would allow specified substitutions of closely related subjects in other departments where possible.”

This overlap rule is confusing but essential to understand.

All Engineering majors and all but one Science major use the maximum 36 units of overlap by “capturing” the REST and Institute Lab requirements. What this means is that those majors specify specific REST and Lab subjects in order to either enable their majors to expand beyond 12.5 subjects, up to 15.5 subjects, or allow their students to take more unrestricted electives of their 180 units required beyond the GIRs. This is not a new phenomenon— overlap rules have existed for decades (we spot-examined the 1980/81 MIT Bulletin and found overlap rules back then).

As a specific example, Mechanical Engineering (Course 2) uses 2.001 and 18.03 as their specified REST classes and 2.671 as their specified Institute Lab subject, as shown in Fig 1. As another example, Physics uses 8.03/8.04 and 18.03 as their REST subjects, and 8.13 as their Institute Lab subject. 

Fig 1. Illustrative example of overlap between SME GIRs and Course 2, where the REST subjects are specified as 2.001 and 18.03. Every engineering major and all but one science major specify REST subjects.

If the REST requirement is removed, what happens to those classes? If those REST units are repurposed for a different set of required GIR classes (like computation or PSM), then the existing classes that use REST no longer fit in the major (Fig 2). The only way to keep those classes in the major would be for Mechanical Engineering to not require 2.001 or 18.03 (or slot those in and remove two other required subjects from their major). This situation is not unique to Mechanical Engineering—many majors do not have 24 free units to incorporate the REST subjects. 

Fig 2. Removing the REST requirement and repurposing those units for a new requirement (like computing, PSM) means that the classes that used to be REST subjects (2.001, 18.03) have no place in a full major. 

But why not remove the REST requirement and not return those units to the major, in other words, reduce the maximum major size? During our listening tour, we heard from several engineering departments that they felt disadvantaged with the current situation; i.e., 15.5 subjects (including 3 subjects of overlap) was not enough and left their graduates at a competitive disadvantage with peer institutions that allow larger major sizes. Performing some historical analysis, we found that engineering majors have roughly stayed the same size over at least the past 40 years or so, suggesting that the current size of engineering majors is not due to some recent expansion (which we could arguably push back on). We thus decided to pursue approaches that would not force a reduction in the maximum size of majors.

Thus, we have recommended that, in removing the REST requirement, the units be returned to the majors. This doesn’t make the majors any larger than they used to be (Fig 3), but it highlights one of our design constraints, which is that we wanted to avoid designs that required majors to remove discipline-specific subjects in order to accommodate the new set of GIR requirements. We thus went through all the majors at or near the units limit (see Appendix F, p. 97) to ensure that there would be room to add back in requirements that were reduced or eliminated (for example, physics and several engineering majors would likely want to require 8.02 and thus add it into their major). 

Fig 3. In TFUAP’s proposal, we remove REST and return the units to the majors. This does not increase the size of majors but allows those REST subjects (2.001, 18.03 in this case) to remain within the majors.

Returning the REST units to the majors prevents “breaking” a large number of existing majors. But what this does mean is that removing the REST requirement does not free up 24 units. 

This leads to the last suggested approach for finding units, which is to incorporate a requirement (computing, PSM most often suggested) into the majors. 

Here we need to consider when it makes sense to have a GIR subject outside the major versus inside the major (via overlap). If a required subject already exists within all majors, independent of the GIRs, then “capturing” it by renaming it into a GIR (and adding overlap units) is possible. This approach does not free up units (because that subject already takes up units in the academic program), but it does ensure that this subject would always remain within the academic program. As a somewhat fanciful but illustrative example, we could place the entire current set of SME GIRs (8.01, 8.02, 18.01, etc.) inside the majors and increase the overlap allowance accordingly (Fig 4). No units are liberated in the process.

Fig 4. Incorporating the entire current set of GIRs into the majors with overlap does not free up units, shown here using Mechanical Engineering (Course 2) as an example. 

None of the new requirements proposed by TFUAP (computing, PSM, or TI) exist in all majors; they cannot simply be captured and renamed into a GIR. Thus, in proposing them, we had to create room by reducing or eliminating current requirements (such as 8.02, bio, and chem) and introducing flexibility in added requirements (compute, PSM). 

Requirements that exist as a list of classes that are not necessarily interchangeable are good candidates for inclusion with overlap. REST and the Institute Lab in the current SME GIRs fall into this category, and are the reason they are most often used as overlap units in our current system. In our proposal, Probability, Statistics, and Machine Learning, and Teamwork-Intensive subjects fall into this category. This is the reason we proposed to keep 12 units of potential overlap and apply them to these requirements. This design has the feature that it keeps the maximum major size at 15.5 subjects (with 12 units of overlap). In our analysis, this design will not “break” any majors.

Importantly, we did not restrict ourselves to solutions that required no changes to the majors, because the majors have evolved in response to the current set of GIR requirements, and almost any change in the GIR requirements will require changes to a number of majors. Setting a constraint of “no change” to the majors would effectively eliminate any change to the GIRs.

Overall, it is this balancing act that led us to the current TFUAP proposal. We appreciate that others in the community might use a different set of design constraints, or may come to a somewhat different proposal using the same set of constraints. We hope that the community understands that our goal was to balance a diverse set of priorities to come up with a solution that best serves our students. We close with a prescient quote from the Silbey report (pg. 52):

In closing, however, it is important to note the following:

After two years of considering this matter, what is not in serious widespread dispute is whether we will move toward greater flexibility in our common science and engineering requirement or whether we will open up the requirement to include new areas such as computation, engineering, and complex systems. The real question is how to do this in the best interests of MIT students, considered as a whole.

Twenty years later, with an entirely different set of people, TFUAP has reached a similar conclusion, namely that we need to incorporate flexibility in our Science core in order to incorporate new areas.

SMC GIRs

Why add a Computation GIR when most undergraduates are taking computation already?

We know that most undergraduates already take an introductory computation subject. While most do so to satisfy a major requirement, some choose to do so even if not required, whether to explore a new topic, develop what they consider necessary literacy, or for another reason. Adding a computation GIR ensures that all students have this necessary literacy and develop it early enough that every major has the opportunity to leverage this background. TFUAP has heard from instructors in majors without a formal computation requirement who have to spend time teaching basic computation skills in case some students lack prior experience. In addition, the fact that most MIT undergraduates already take an introductory computation subject means that computation is a hidden requirement for students. TFUAP’s conclusion was that we should ‘call a spade a spade’ and make it an actual requirement.

Could we remove 18.01 from the GIRs to make room for additional content? If students do not or cannot take calculus in high school, couldn’t we offer a summer boot camp or online class to get them up to speed?

Based on discussions with MIT Admissions and FLI students, this change would have serious and negative impacts on MIT’s ability to recruit talent from less-resourced areas of the country. Many high schools do not offer calculus, and students may not have access to affordable community college or online offerings. TFUAP maintains that 18.01 should continue to be a GIR to serve the many talented students who lacked an opportunity to study calculus before arriving on campus.

Does eliminating the 8.02 requirement and allowing for 6 units each of Chemistry and Biology signal that MIT considers science to be less important than Computing?

No, we did not wish to signal this. Several key points:

1. Physics and Math are being placed in a different category than Computing, where students will be required to take more units of these than Computing.
2. Chemistry and Biology are still required, and we directly discussed their importance in the proposal. We decided Chemistry and Biology should be treated as equal to Computing and PSM, with students given the choice of what to emphasize in their curriculum, and majors the ability to specify what is appropriate for their discipline.
3. The proposal recognizes the fact that modern science is being driven by the latest technology – including Computing and PSM – and provides opportunities for faculty to teach these disciplines in ways that utilize multiple disciplinary perspectives.
4. Innovative scientific discoveries are being driven by interconnections between fields.  The proposal re-envisions what a modern science education should be; it doesn’t have to be the siloed disciplines that are related to what students see in high school.

Flexible Foundations

How can a 6-unit class be sufficient to teach more than surface-level content in any of the listed disciplines?

While a 6-unit subject may not cover every introductory topic in sufficient depth to serve as a prerequisite for certain majors, it can cover the many essential topics in a given field that every MIT graduate, regardless of major, should know (i.e., essential literacy) and also provide creative analytical ways of thinking. Decisions about what subjects satisfy the SMC GIRs will fall to the proposed subcommittees, which will include representatives from the GIRs’ home departments as well as downstream departments that utilize the GIRs (and overseen by CUP). TFUAP notes that 6.100A Introduction to Computer Science is an existing example of a 6-unit subject that provides common literacy in computing that every MIT graduate,  regardless of major, should know, and 6-unit subjects that may already satisfy PSM also exist.  We also recommend that 6-unit subjects be offered for the entire semester, as this format of subject offering tends to work better than a half-semester 6-unit subject. 

Will departments need to offer “catch-up classes” for students who take a 6-unit GIR and then decide to major in something that requires 12 units of that GIR? How will departments maintain all the different types of GIR offerings expected by the proposal?

Analysis using data from Institutional Research indicates that it is relatively rare for students to completely shift fields. We conducted this analysis based on data showing the intended majors students listed on their applications and what majors they ultimately declared. While many students change their minds, they most often shift between related fields with similar prerequisites (e.g., from 7 to 10B). Our analysis estimates the pool of students needing to “catch up” as ~40 students per year per subject, and that is before any behavior change that might be expected under the new system. Thus, in our final report, we will not propose that departments offer catch-up subjects. Instead, we will propose that departments permit students to take ASE exams for 12-unit GIRs if they have already completed a 6-unit or integrated version, and that MIT should track the number of catch-up students to see if the numbers bear out.

Truly undecided students might be advised to take the 12-unit versions of any subject needed for majors they are considering, which may result in them taking seven 12-unit subjects to satisfy their GIRs. Alternatively, students can take a 6-unit version, knowing that they may need to take a 12-unit version later if they change their intended major. 

In terms of departmental offerings, departments offering flexible foundations GIRs would need to offer two versions: a 12U and a 6U version. There is no obligation to offer integrated versions of subjects, though we encourage the community to develop them. Because the total number of students to be taught remains the same, the existing 12U versions of subjects would decrease in size, freeing up resources to offer the 6U versions, and integrated versions if desired. Resources will be needed to develop new subjects, but in a steady state, the total enrollment across 12U and 6U subjects will equal the current enrollment of the 12U subjects.

Why not put the Probability, Statistics, and Machine Learning (PSM) requirement in students’ majors rather than in the SMC GIRs? Don’t these concepts make more sense in the context of one’s field?

Majors that feel strongly about teaching PSM a certain way (including both basic and applied concepts) can choose to use their 12 units of specified GIRs (‘overlap’) to prescribe a particular version, including a sophomore or junior-level major class. That said, many students and majors already require probability, statistics, or machine learning classes that are not discipline-specific, such as 18.05, 6.3700, 6.3800, and 6.3900.  Thus, we disagree that PSM must be discipline-specific in order to be valuable. 

TFUAP also proposes that the ad hoc subcommittee tasked with implementing PSM consider rules for when two major classes could be combined to satisfy PSM. Because PSM serves both a foundational role (in some, if not most majors) and an essential literacy role (given the importance of data literacy when interpreting information in contemporary society), TFUAP believes that at least part of the content in PSM subjects should be generally relevant concepts rather than solely applicable to major-specific contexts. 

Students don’t need to learn to write computer programs now that everyone is using large language models to code. Isn’t your recommendation of a Computation GIR obsolete?

We believe that this conflates computational thinking with the generation of program code. Saying “AI will take over programming” is like saying “digital calculators took over math.” Digital calculators certainly do most of our arithmetic these days, but we still teach students how to do arithmetic because understanding is important to advance to higher math. Software like Mathematica and Wolfram Alpha has been used to solve algebra and calculus problems for decades, but we still teach students how to do that math for themselves. AI can solve an increasing number of science problems and generate text in a variety of styles, but we still believe that students should have science literacy and scientific ‘ways of thinking’, and learn how to structure an argument and communicate. AI will likely take over the writing of straightforward code, but students will still need computational thinking: understanding how to translate a problem so it can be solved computationally, and how to recognize and debug incorrect algorithms.

Why did you change the recommended pacing requirement for SMC from 2 years to 5 semesters? 

The SMC pacing requirement is intended to ensure that students take SMC subjects early enough in their time at MIT that the classes have a chance to be influential in the students’ academic path and can serve as a foundation for advanced work where appropriate. The pacing requirement also serves a cohort-building role, giving students a shared early experience.  Some students, meanwhile, pointed out that requiring 72 units of SMC along with two CI subjects (another paced requirement) and accounting for the desired 1 HASS subject each term (which will overlap with CI subjects) constrains 10 of the typically 16 subjects they would take in the first two years, preventing them from fully exploring their interests. Thus, we felt that extending the pacing to 5 semesters would strike a better balance between these two pressures.

Lab

Why are you getting rid of the Institute Lab requirement? Aren’t hands-on labs an essential part of an MIT education and an inspiring opportunity that all students deserve to have? Why not strengthen the lab requirement instead?

TFUAP recognized the importance of hands-on work in our proposal. Given that the student academic experience is completely full, we decided not to introduce another requirement, but to recognize students who substantially participate in these activities. 

To the extent that community members desire a hands-on laboratory experience, the Institute Lab, as construed for many decades, does not do that and is not currently the mechanism to achieve that (see history below). Hands-on work is in the lifeblood of the MIT undergraduate experience, even though the Institute Lab does not require students to do so. Currently, majors enrolling at least 40% of our undergraduates include Institute Lab subjects that do not actually entail hands-on laboratory activity. Indeed, hands-on making and breaking, almost uniquely among educational experiences (and certainly different than Math, Physics, etc.), can be done extra-curricularly, via clubs, personal projects, UROPs, internships, and so on.

Given all this, within TFUAP, we decided to highlight subjects and experiences that incorporate hands-on work through a new icon that can be incorporated into the Subject Listing, a club’s website, and so on.  This will highlight the large amount of making and breaking that occurs at MIT, and allow advisors to direct students to relevant experiences.

The Lab requirement has never actually mandated hands-on experimentation. The Institute Laboratory Requirement (Lab) originated as an outcome of the Zacharias report of 1964. At some point (we are unsure when it began, but the wording was removed in AY2015/16), the Committee on Curricula guidelines specified that Lab subjects should study “phenomena of the natural world”, which many have interpreted as requiring hands-on experimentation. However, 15.301 (and its precursors, first introduced sometime in the 1970s), is a social science Lab subject that does not involve experimental apparatus. In 2006, CoC approved 18.821, which uses computational methods, as a Lab. 

One could argue that instead of removing the Institute Lab requirement, we should modify or fix it to mandate hands-on experimentation. We decided against this approach for a number of reasons. First,  hands-on work is a vibrant cultural aspect of the undergraduate experience; it is already working and not in dire need of adjustment. Second, because hands-on work can occur extra-curricularly, mandating a curricular requirement forces some students to do extra, potentially removing some of the joy they feel in this. Third, the reason the Institute Lab does not mandate hands-on work is because different constituencies have deeply held differing views on what the requirement should provide.

In the many majors that rely on hands-on experiences, TFUAP expects that such subjects, whether conforming to current lab GIR standards or not, will continue to be required of all their undergraduates. It is simply inconceivable that disciplines like Electrical Engineering, Mechanical Engineering, or many science disciplines will remove hands-on laboratories because the Institute does not require it (as one example, the iconic 2.009 subject is not an Institute Lab). 

HASS GIRs

Wouldn’t the Moral and Civic Perspectives (MCP) requirement make more sense in the majors so students can study ethical dilemmas specific to their field?

Ideally, majors would reinforce and apply MCP concepts in their required coursework, but TFUAP feels that a strong introduction to MCP should come from departments with expertise on those topics, namely the departments that already teach HASS subjects. We are concerned that MCP subjects taught solely by science and engineering departments would lack sufficient scholarly rigor and depth, particularly if such subjects are also focused on technical content. That said, we would welcome subjects that are co-taught by HASS and STEM faculty that cover MCP subjects both broadly and in major-specific contexts. We also note that students are members of society independent of their professions, and should be prepared to consider moral and civic perspectives in their personal lives as well as their professional lives. 

How will it be possible to scale the Moral and Civic Perspectives (MCP) requirement up so that there is sufficient capacity for every MIT undergraduate to take it?

This is an important point, as MCP class sections are going to need to be limited in size in order to foster discussion. As a starting point, TFUAP worked with SHASS to compile a list of courses that could conceivably be counted as MCP courses with some or little modification (though this is ultimately up to the proposed MCP subcommittee to decide).  From these data, it seems reasonable to expect that there would be ~ 20 subjects across SHASS that would be counted and that at least some of these subjects could have multiple sections. Thus, it seems feasible, though by no means trivial, to serve ~1100 students/year.

Impact on Majors 

How does the proposal impact the size of majors, what GIRs majors can rely on as prerequisites, and how students can double-count major requirements to satisfy their GIRs?

There are three aspects of the proposal that have to do with some aspect of how GIRs and majors intersect: 

1. Specifying Flexible Foundations: Within the SMC GIRs, TFUAP is proposing a category of four subject areas split across 36 units in Flexible Foundations. All majors can assume students have 12 units of background in each of the three Common Foundations subjects and 6 units of background in each of the four Flexible Foundations subjects. In addition, each major can mandate that students take 12-unit versions of up to two of the four Flexible Foundations topics. These expectations will be communicated to first-year students so they can select their GIRs based on the majors they are considering. 
2. Maximum Major Size: Under the current system, majors can require up to 12.5 subjects and specify an additional 36 units (most commonly REST and Lab) within the GIRs. Under TFUAP’s proposal, majors could require up to 14.5 subjects and specify an additional 12 units within the GIRs, which could include any of the HASS, Teamwork-Intensive, or PSM (Probability, Statistics, and Machine Learning) requirements. In both cases, the effective maximum number of units that a major could expect of students (including requirements and specified GIRs) would be 186 units (15.5 standard subjects).
3. Abolishing Non-Overlap Guideline: While TFUAP suggests that a cap on the number of GIRs that majors can specify be maintained as above, TFUAP proposes that there be no limit to the number of subjects that students may choose to count for both major requirements and GIRs. For example, a student who majors in a HASS field would be allowed to concentrate in the same field, regardless of whether the HASS major is their “primary” or “secondary” major, as long as they satisfy the requirements for both the major and the concentration.

Policies

Part of what makes MIT so special is that students have the flexibility to take risks and challenge themselves beyond what another university might allow. This includes things like “double-booking” or taking two classes that meet at the same time. Why would you get rid of this aspect of MIT culture?

While TFUAP values student choice and balanced student freedom with shared goals for learning in our design, double-booking is a case where we believe instructors deserve more of a say. Instructors have shared concerns about double-booking, noting negative impacts on the classroom environment and student learning, as well as extra work required to run makeup exams or otherwise help students compensate for missing class. Instructors also have concerns that students are spreading themselves too thin and not becoming as deeply engaged in their studies as they should. Currently, many instructors are unaware that a student in their class is double-booked until the student asks for such support, and instructors vary in their ability and willingness to accommodate such requests. Our policy does not “ban” double-booking outright, but instead implements a process by which students can double-book certain classes but not others and petition to double-book based on extenuating circumstances. We believe this approach balances the needs of students and instructors and will result in more transparency for both sides.