Extracting Year (only) information from Non-Standard Dates

dates <- c("January 12, 1988", "March 4, 2006", "1958")
gsub(".*, *(.*)", "\\1", dates) # notice the it is comma blank asterisk in the string

I have some data whose date column consists of two types of date entries:
(a) year-only entries (eg "1983") and
(b) full date info (eg September 12, 1962).

Here's what the non-standard date info looks like:

mode(non.standard.dates)

[1] "numeric"

head(non.standard.dates)

[1] July 15, 1925     February 13, 1923 July 10, 1988     February 24, 1931
[5] 1952     January 9, 1957
1638 Levels: 1732 1735 1736 1740 1745 1748 1749 1750 1752 1754 1757 1760 ... September 9, 1898

I just want to extract out the year information (viz in the example above: 1925, 1923, 1988, 1931, 1952 and 1957) and create a "year" column. Have tried using the R date libraries without success, but they seem more designed for working with dates which are standardized the same way.... I'd really appreciate suggestions from more experienced R folks. Thank you.