# Mike Hammond's tutorial exercise on searching the Brown Corpus 
# for words that begin with /b, d/ versus /n, m/ ...
setwd('C:/Documents and Settings/install215/®à­±/HammondTutorial')
x=read.table("bc.txt",as.is=TRUE)
dim(x)
# [1] 1021281       1
names(x)="fact"
x$char=as.character(x$fact)
stopVnasal=c(NA,NA)
stopVnasal[1]=length(grep("^[bdBG]",x$char,value=FALSE))
# [1] 75943
stopVnasal[2]=length(grep("^([nmNM]|[kK]n)",x$char,value=FALSE))
# [1] 63629
names(stopVnasal)=c("stops","nasals")
windows(height=3,width=4,pointsize=12)
par(family="serif",oma=rep(0,4),mar=c(0))
barplot(stopVnasal,ylim=c(0,80000))
box()
chisq.test(stopVnasal)
# X-squared = 1086.426, df = 1, p-value < 2.2e-16