closure.R

library(purrr)

foo <- function(x) {
  return(function(y) {
    y + x
  })
}

args <- list(1, 2)
foos_map <- map(args, foo)
foos_lapply <- lapply(args, foo)

foos_map[[1]](1)
# returns 3
foos_lapply[[1]](1)
# returns 2

foos_map[[1]](2)
# returns 4
foos_lapply[[1]](2)
# returns 3

I've synthesized what everyone has said on this issue into a writeup that works through the problem in steps here: http://www.win-vector.com/blog/2017/02/iteration-and-closures-in-r/

This is related to metaprogramming. Here is an old blog post, discussing a similar issue of generating lots of functions having similar templates: Metaprogramming in R with an example: Beating lazy evaluation. One way to resolve this is to use substitute. It is a bit verbose though without using purrr.