18 December 2009
[Updated on 20 December, see note at end.]
Famously, Google has an informal motto of “don’t be evil.” It is a dig at the business practices of Microsoft and an admonition to its employees to use Google’s wealth and market position only in the service of goodness and light. It’s a fun motto, but it’s not quite right; mainly because Microsoft is not really evil. The problem with Microsoft is that they are mediocre; the company produces products that are merely adequate and by force of its wealth and market position squashes technically superior and more innovative products. And if Microsoft can’t beat them, they buy them. Sometimes this quashing of the competition is deliberate, and sometimes it is just inevitable—the giant is so big it can’t help but tread on anyone in its way.
But by focusing on the “evil,” Google has missed the real trap and it is in serious danger of becoming another purveyor of mediocre products, and like the company they are trying not to emulate, a mediocre company so large that it crushes it competitors simply by announcing it is moving into a particular market. The problems with Google Books are legion and well documented. Gmail is the preferred home of automated comment-spammers who are destroying the blogosphere. (Some 80% of the spambots who hit Wordorigins.org each day use Gmail accounts. The problem is so bad that I’ve seriously considered banning Gmail accounts from the site and have only refrained because so many legitimate participants use the service.) These aren’t bad products—they have many good attributes to recommend them—but they have myriad problems and Google seems to be doing little to correct the deficiencies. Mediocrity. And now Google has moved into the dictionary market.
But the new Google Dictionary is not merely mediocre, it is bad. It is laughably bad—Plan Nine from Outer Space bad (only not as funny). If it were from any company other than Google, I’d dismiss it out of hand. But it is Google and if they ever decide to promote it—which fortunately they aren’t doing at the moment; they seem to be doing their level best to hide the fact that it even exists—it has the potential to destroy any number of really good free, online dictionaries that already exist.
So what is wrong with it?
1) A complete lack of explanatory material. If you open up any reputable dictionary (e.g., one from Oxford University Press, Merriam-Webster, American Heritage) and you will see up front a thorough explanation of how the dictionary was assembled, what the editorial practices and standards are, how words are selected for inclusion, what corpus of usage citations is used as a basis for definitions, what abbreviations are used and what they mean, etc. Most people never read this front matter, and the casual user of a dictionary really has no need to, other than maybe consult the list of abbreviations. But the front matter is important. If you are going to rely on a dictionary as an authority, you need to be able to evaluate how good an authority it is. Google Dictionary provides absolutely no explanatory information; Google doesn’t even tell you what the abbreviations they use mean.
2) Unskilled and amateurish definitions. Writing a good, concise definition that encapsulates all relevant uses of a term is an art, a very subtle art. It takes a lot of practice to do it right. Clearly, whoever is writing the definitions for Google is very inexperienced. Take for example the definitions of the verb to dive (which I looked up on a whim because we had recently been discussing its origin on the Wordorigins discussion forum). Google has:
1. If you dive into some water, you jump in head-first with your arms held straight above your head.
Compare this to Merriam-Webster Online (based on the 11th edition of their Collegiate Dictionary):
to plunge into water intentionally and especially headfirst.
Or to the OED 2nd edition:
to descend or plunge into or under water or other liquid. (Usually, unless otherwise stated, to plunge head-foremost.)
Not only is the Google definition verbose (to be fair, the OED definition could also use some pruning), but it is extremely limiting; according to Google you cannot dive into water with your arms at your side, besides, if you are diving head-first aren’t your arms below your head? This is not an isolated instance. The dictionary is replete with shoddy definitions like this one.
3) No consistent editorial standards. Picking a random word from out of my head, I entered supplicant into the dictionary and got this:
supplicant /s’ʌplɪkənt/
Synonyms:
adjective: beseeching, pleading, suppliant
noun: suppliant, petitioner, applicantsupplicants plural
1. A supplicant is a person who prays to God or respectfully asks an important person to help them or to give them something that they want very much. | N-COUNT | formal
● He flung himself down in the flat submissive posture of a mere supplicant.
I then clicked on the link to the synonym suppliant and got this:
supplicant /ˈsʌplɪkənt/ DJ /’sʌplɪkənt/ KK
Synonyms:
adjective: beseeching, pleading, supplicatory, imploring, appealing, supplicant
noun: supplicant, petitioner, suitor1. a person who asks for something in a humble way, especially from God or a powerful person
| noun | formal
Variant: suppliant /ˈsʌpliənt/ DJ /’sʌplɪənt/ KK
Note that this is not suppliant, but a second entry for supplicant, one with a different definition, different phonetic notations, different synonyms, and different editorial notation. (Note, I did not cherry pick this one after going through many entries. Supplicant was literally the second word I looked up, after dive.)
Again , compare Google’s definitions to the OED and Merriam-Webster Online ones. Both these dictionaries, betraying their print roots where space was at a premium, define it as “one who supplicates.” Clicking through the cross-references to the verb finds the definitions. OED:
intr. To beg, pray, or entreat humbly; to present a humble petition
and
trans. To petition humbly.
M-W:
intransitive verb : to make a humble entreaty; especially : to pray to God transitive verb 1 : to ask humbly and earnestly of 2 : to ask for earnestly and humbly.
Editorially, what do the abbreviations DJ and KK mean? I would bet that 99.99% of the people who consult the dictionary would have no idea. I had to look them up; they are different types of phonetic notation named for their inventors, Daniel Jones and Kenyon & Knott. Not only do they not explain it, but Google does not provide consistent phonetic standards across the entries. This combination of amateurism and really arcane lexicographic distinctions is truly bizarre. And what is the difference between noun and n-count? Unlike the phonetic abbreviations, it is not difficult to suss out the meaning of these, but again there is no standardization of editorial practice across the various entries.
Google should either stay out of the dictionary market altogether or go and buy an established dictionary publisher, or at least license their content. Lexicographic publishing is not the most lucrative of businesses, and I’m sure there are several publishers that would be happy to make a deal. It’s not like Google doesn’t have the money to do it right.
In addition to Merriam-Webster Online, if you’re looking for a good, free online dictionary, check out dictionary.com or the new wordnik.com. Dictionary.com (part of Ask.com) is a search engine that delivers professional dictionary entries. You get multiple results from each query. Wordnik.com is a new entry into the online dictionary market. It has licensed American Heritage’s content and provides some results from older, public domain dictionaries. But it also has features to incorporate user input and feedback. I’m guessing that it will provide some really neat and innovative features in the future. (I have no inside knowledge about this, just my speculation.)
Google Dictionary also includes bilingual features for the most common languages. I have not looked at these extensively, although I have a posting in the works on what makes a good online, bilingual dictionary.
(Google has for a long time maintained a very useful feature in its search engine. Simply by typing “define: _____” into the search box, you get a list of definitions from various web sites for the term you enter. This is a godsend for finding the meaning of jargon and abbreviations. (Although it didn’t help with either DJ or KK.) To their credit, it seems that Google is continuing to support this feature.)
[Note, 20 Dec 2009: It has been pointed out to me that Google has indeed obtained their dictionary content through licensing. They are using the Collins English COBUILD dictionary. Many of the issues that I have pointed out are inherent in the COBUILD data and not the creation of Google. This, however, does not change my criticism. Google made a very poor choice in which data it chose to license and as publisher they bear the responsibility.]