Tomlin disagreed with the samples of languages used in the previously published distributions of word orders. He attempted to choose a representative sample. There is of course major problem of what is meant by representative. There are perhaps six thousand extant languages. Would a representative sample treat a language having only a few thousand speakers the same as one having a billion or so speaker? Another concept of a distribution would be based upon the total speakers. The basis for the selection of the sample might be to estimate what is the probability of a randomly selected speaker speaking a language with the SVO word order or any of the other possibilities? Tomlin's sample seems to be some compromise between treating any language as being a legitimate linguistic observation and weighting the sample toward the major languages in terms of the number of speakers.

In the above table the square brackets around Verb and Object reflects that usually these two are bound into a substructure.

The salient result is that most languages are either SOV or SVO. Nevertheless there are working languages using the most of the other word orders. An example of a OVS language is Hixkaryana, a language spoken in Amazonia by a Carib tribal group. There is some indication that a language called Warao spoken in Venezuela is of the OSV type.

Tomlin gives a cogent explanation for the distribution of languages among the word order types. He enunciates three principles which go a long ways toward explaining the relative abundance of the word order type.

• Theme First
• Verb-Object Bonding
• Animated First

(To be continued.)

Estimates of the Word Order Distibution of Languages