nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpqez3yya8tpgge7lk4jq2zxz0cj3d2mcgev9mffd4ec7tfy84hre4skst6k2 (nprofile…t6k2) But most types are given universal properties precisely by reflection. Terms of sigma types are in bijective correspondence with dependent pairs of terms. If we are working in an LF, terms of pi types are in bijective correspondence with meta functions!
Introducing equality reflection as a rule in MLTT in order to make extensional type theory, if MLTT is specified as something sensible like a GAT or SOGAT, seems like a perfectly reasonable thing to do! We're already quotienting the syntax by eta and beta for all the rest of the type formers, why not equality too?
Owen Lynch
npub1cj…fxs07
2026-01-26 20:53:25 UTC
in reply to nevent1q…2z3h
Author Public Key
npub1cjyyl7y66997ewku92pzvvhkyec43k7cde44rhgayxdwwl63axdqafxs07Seen on
wss://relay.ditto.pubPublished at
2026-01-26 20:53:25 UTCEvent JSON
{
"id": "eee1eacbfd41ca81fa5fcfe03e66f659fe36f49cb13ffd76af810d0a1160327a",
"pubkey": "c4884ff89ad14becbadc2a822632f6267158dbd86e6b51dd1d219ae77f51e99a",
"created_at": 1769460805,
"kind": 1,
"tags": [
[
"p",
"c8a24274eb0a119f7ed590142309f8945aade119617694b6b9c796921eb71e6b",
"wss://relay.ditto.pub"
],
[
"p",
"5c675fe7561c52f2c0ea8db391d2c3a46556bec9a3c0a4b4f10f234fe03f58ca",
"wss://relay.ditto.pub"
],
[
"e",
"49b03272ea22d3b8035574e5f445e767f7fac973989151fa5c8a4ebe948ab97d",
"wss://relay.ditto.pub",
"reply"
],
[
"proxy",
"https://mathstodon.xyz/users/olynch/statuses/115963383375720700",
"activitypub"
],
[
"client",
"Mostr",
"31990:6be38f8c63df7dbf84db7ec4a6e6fbbd8d19dca3b980efad18585c46f04b26f9:mostr",
"wss://relay.ditto.pub"
]
],
"content": "nostr:nprofile1qy2hwumn8ghj7un9d3shjtnyd968gmewwp6kyqpqez3yya8tpgge7lk4jq2zxz0cj3d2mcgev9mffd4ec7tfy84hre4skst6k2 But most types are given universal properties precisely by reflection. Terms of sigma types are in bijective correspondence with dependent pairs of terms. If we are working in an LF, terms of pi types are in bijective correspondence with meta functions!\n\nIntroducing equality reflection as a rule in MLTT in order to make extensional type theory, if MLTT is specified as something sensible like a GAT or SOGAT, seems like a perfectly reasonable thing to do! We're already quotienting the syntax by eta and beta for all the rest of the type formers, why not equality too?",
"sig": "a14690d17671d9629ce88ed4de1b88a86de76dd36e06cb37a5467fed37f09f07c71c14e1c17fff2f645d34a47e955295cca2d8b8378e396f59a8330f02d94a64"
}