Sample Sentences in Thai Dictionaries

Sample Sentences in Thai Dictionaries…

While reviewing Thai dictionary apps for my series, iPhone, iPad and iPod Touch Apps, I came across a subject that needed clarification: sample sentences in dictionaries.

As 
Chris Pirazzi already answered my question about the LEXiTRON dictionary database (developed by NECTEC) in a previous post, Backstage View into the Process of Creating a Thai Dictionary, I knew he was the person to ask.

So before I post my review, please take the time to read Chris’ valuable insights into sentence samples found in Thai dictionaries.

Chris Pirazzi on Sample Sentences in Thai Dictionaries…

I’m really glad you mentioned sample sentences. We may indeed add sample sentences at some point, but sample sentences are an area where there is a LOT of misunderstanding and mis-set customer expectations, and you could really help a lot of people in your review by helping to head off these mis-set expectations before the customer gets disappointed.

First of all, customers should be aware that the quality of sample sentences varies wildly between apps, so customers should be sure to look at more than just quantity. Most apps actually get their sample sentences by having a computer program crawl through huge, freely available, error-ridden bilingual datasets available on the internet with no human intervention or editing. In many cases, the sample sentences contain errors, or even more often, they do not even demonstrate the meaning of the word being defined at all. For example, an entry for “หก” with the English translation “The number 6″ may contain many sample sentences for the other meaning of หก, “to spill (a liquid).” The sample sentences in these cases are at best confusing and possibly misleading. With some apps (and I never could figure out why) you will even see sample sentences for a Thai word that don’t even contain the Thai word! Typically, if an app boasts a huge number of sample sentences (like, tens of thousands or more), that is a major red flag that the sentences are crap. Doing sample sentences properly requires humans to edit the sentences of each entry for relevance, and that takes almost as much work as creating the dictionary dataset in the first place. Almost no vendor is willing to take this time for editing.

Secondly, and even more importantly, challenge your reader to ask why they want sample sentences.  There may be other ways of getting what they want that are simpler and more direct. Let me explain.

Sample sentences are a little like transcription: at first, when looking for Thai learning materials, Thai learners always ask for a transcription system that is as “English-Like” as possible, and they may even choose their app by that criterion. It isn’t until much later that they realize that due to the unavoidable reality of Thai language and how its sounds differ from English, the goal of being “English-like” is not only impossible but it may actually damage their ability to learn Thai sounds properly (e.g., transcription systems which over-simplify Thai sounds so that ส้ม and ซ่อม are written the same way, or เป็ด and เผ็ด are written the same way), or at the very least the goal of being “English-Like” may actually make the transcription system more complicated and make it more hard for them to learn Thai than they could with other systems. I talk about that at length in Slice of Thai: Pronunciation Guide Systems for Thai.

At first, customers also ask for sample sentences, but sample sentences can lead the customer to a similar dilemma. When we were beginning our multi-year dictionary production process, we asked ourselves why it is that people ask for sample sentences. The answer is that it helps give more information about a given translation, for example:

• for a given translation from language 1 to language 2, which SENSE of the word is being translated?  For example, if there is an English entry for “glass” that shows a Thai word, then is that the Thai word for “drinking glass” or the word for “pane of glass?”
• what prepositions and other linking words need to be used along with a given word?  For example, when I want to say “wait for him”, I can see that there is a Thai word “รอ,” but what (if any) preposition should I put in in place of “for”?
• what level of formality (e.g. slang, formal) does the word have?
• what are the word’s classifiers, if it is a noun?

Typically, bilingual dictionaries will try to answer these questions by providing sample sentences. 

But even if the sample sentences are carefully hand-crafted and hand-edited by humans (and so far I have never seen an iOS/Android app where they are), sample sentences are a very poor way to answer the questions above, because the reader has to read the sentence, understand its parts, and then think backwards to get the answer to the original question they really wanted answered.

We decided that it’s much better to spend our effort answering the important questions for the user directly. We are the only Thai-English English-Thai dictionary that we are aware of that was designed from the ground up to help English speakers who are learning Thai in this way.

When giving definitions, we provide glosses to clarify shades of meaning (e.g. “glass (drinking)” vs. “glass (pane)”), as shown on our website at Designed for English Speakers.

We have specifically designed our headwords to solve the preposition/extra word problem. For example, we include a transitive verb entry “wait for” that translates to “รอ,” and this is a specific, explicit cue to the user that they do not need to insert a Thai word corresponding to “for” when using the Thai verb รอ. We talk more about how this works in our application Help under “Speaking and Listening” then “Verbs, Objects, and Prepositions.”

We specifically notate the register (slang, formal, …) of each word using symbols, rather than trying to make the user guess from sample sentences. You can click on “Word Register” in our app Help to get the details.

And of course we explicitly notate classifiers too.

There are still cases where sample sentences can be handy, but we feel we’ve delivered a much, much greater bang for the buck by spending our finite development time by going right for the information that Thai learners need.  We may still add sample sentences as well. No matter what, we will continue to listen to our customers’ requests for what information they want in each entry and provide that in the most direct and useful form we can.

Chris Pirazzi,
Word in the Hand: Talking Thai-English-Thai Dictionary | Slice of Thai | Thailand Fever