locus {emuR}

Calculate locus equations for two-dimensional data

Description

The function plots a locus equation and returns associated statistical information.

Usage

locus(
  target,
  onset,
  labels.vow = NULL,
  yxline = TRUE,
  plotgraph = TRUE,
  axes = TRUE,
  ...
)

Arguments

target

a numerical vector typically of F2 values at the vowel target

onset

a numerical vector typically of the same length as target of F2 values at the vowel onset

labels.vow

an optionally character vector for plotting labels at the points (target, onset) of the same length as target

yxline

optionally plot the line target = onset. Defaults to True.

plotgraph

a logical vector for specifying whether the data should be plotted. Defaults to True.

axes

A logical vector indicating whether the axes should be plotted

...

graphical options par

Details

A locus equation is a straight line regression fitted with lm() in which the F2- values typically at the vowel onset are regressed on those of the target. The slope can be used to give an indication of target-on-onset coarticulatory influences.

The best estimate of the locus frequency is where the locus equation bisects the line target = onset.

Value

A list containing regression diagnostics of the function lm() that can be accessed with summary() and the estimated locus frequency in $locus. A plot of values in the onset x target plane with superimposed locus equation and line onset=target.

Author(s)

Jonathan Harrington

Examples

 # calculate an F2-locus equation for initial [d] 
# preceding lax vowels produced by female speaker "68".
# the onset is taken at the vowel onset; the
# vowel target is taken at the vowel's temporal midpoint.

# identify initial "d" of speaker "68"
temp <- vowlax.left == "d" & vowlax.spkr == "68"
# get the F2 value at the vowel's temporal midpoint
targ <- dcut(vowlax.fdat[temp,2], .5, prop=TRUE)
# F2 value at the vowel's acoustic onset.
on <- dcut(vowlax.fdat[temp,2], 0, prop=TRUE)

# locus equation plot
result <- locus(targ, on, vowlax.l[temp])
# statistical diagnostics of the regression line (locus equation)
summary(result)
# intercept and slope
result$coeff
# best estimate of the locus frequency, i.e. the
# point of bisection of on = TRUEarg with the regression line
result$locus