Let me be a devil’s advocate against your devil’s advocate: Pedestrian, bicycle and even bus-only crossings costs a fraction to grade separate compared to car crossings. It is much easier for cars to take a significant detours then other transport modes. It is always a much cheaper option to simply close the road for cars and force them to cross at the nearest crossing that is already grade separated, and build a ped/bike-only bridge/tunnel over the tracks.

Cars are the reason this negative externality is so expensive to mitigate. Why should funds from other modes be diverted to it?

> Why should funds from other modes be diverted to it?

If the city's been built and is already car dependent, wishing away the existing uses and the massive amounts of capital dependent on them isn't an option. That is, we have to consider costs at the margin from where we are now.

Since you answered my second order devil’s advocate, I should answer your first order devil’s advocate:

> Why should rail not bear some of the social costs in addition to its direct costs?

The social costs created by private cars dwarfs those created by train, and is seldomly actually mitigated with funds diverted from car infrastructure. Historically this has been an unpopular political choice, made without considering the local communities. Now California is prioritizing a different mode with the hope of reducing car dependency. Why shouldn’t car infrastructure now bear the defunding that other modes have historically suffered in order to accommodate cars historically, infrastructures which has historically created social costs to local communities?

I see this as a way to fix historic wrongs. Rail does not need to bear the social costs of its infrastructure because car owes us a bunch. We should collect on those debts owed by cars.

> Why shouldn’t car infrastructure now bear the defunding that other modes have historically suffered in order to accommodate cars historically, infrastructures which has historically created social costs to local communities?

Because it devalues a whole lot of land and capital that have seen substantial investment under the prior equilibrium.

Fixing historic wrongs or not, churning our transport infrastructure incurs a whole lot of external costs and can't be just considered in terms of "road costs" vs. "rail costs".

I’m confused, are you still playing the devil’s advocate, or do you actually hold this believe?

In the case of the former, then I’ve already answered why it is totally fair for car infrastructure to pay for social externalities of other modes meant to relief car traffic.

In the case of the latter, I don’t know what to tell you except that you are wrong. I didn’t mean defunding as in let existing things churn, I meant removing car lanes and forcing detours so that train infrastructure can be cheaper. The overall effect is better transport for everybody except cars (which already have it plenty good).

Also are you sure that removing car lanes actually devalues land and capital. I’m not so sure that is true. And even if it was, good. Housing in the Bay area is plenty expensive as it is, if cheaper train infrastructure means cheaper housing, then I’d say we’ve succeeded on two fronts.

I hold a pragmatic set of views. I would like us to become less road dependent, but I don't think you can ignore the degree of investment that has been built on the existing road system. Breaking and devaluing existing uses is a social cost incurred by a change in modes of transport.