I am trying to read many files 

I do this and I get the error 

file.list <- list.files(pattern='*.txt')​
dge <- readDGE(file.list, header=FALSE)
Meta tags detected: __no_feature, __ambiguous, __too_low_aQual, __not_aligned, __alignment_not_unique
Error in colSums(x$counts) : 'x' must be numeric

The error is telling you that at least one of the files has a non-numeric character in the second column.

Using list.files(pattern='*.txt') is convenient but dangerous. It is dangerous because it can easily include files that are not htseq output files and because, even if it selects the right files, you don't have any control over the order of files.

In this case, you might have one or more .txt files in the directory that are not a htseq-count output files. Have you checked that?