In written language there are rules to explain away most uncertainties. When there are no explicit rules, there is context. But without a rule and without context, what are we to do?
According to quantum physics, a card perfectly balanced on its edge will by symmetry fall down in both directions at once.
Is it spelled Advisor? Or is it Adviser? Or is it both?
Let’s take a data-journey for one word with two spellings.
In English, an agent noun is how a verb is transformed into the thing doing that action. For example, a swimmer swims and a player plays. To convert a verb into an agent noun, it’s typical to add -er. But -or works, too, primarily for words derived from Latin. And -ist is sometimes used, often for verbs ending in -ize (e.g. apologist).
Originally, adviser was the one true spelling. It was only recently, at the start of the 20th century, that advisor began its ascent, and it will be only a few more years into the 21st before it overtakes adviser.
Interestingly, some early instances of advisor are due to mistakes in reading the scanned pages, called OCR (optical character recognition). In this 1869 case it could have been the hand-drawn “firewood” typeface that confused the algorithm.
It’s often claimed that American English writers use advisor while British English writers use adviser—and the data below seems to agree. Counting instances in American English, advisor (blue) and adviser (red) are about equal in 2008, the most recent year that is available. In British English, advisor (green) is gaining in popularity but adviser (orange) is still preferred. Seems straightforward, right?
But wait, there’s more!
Advisor in the States is far from the default—the NY Times and Associated Press insist on adviser, while the Federal Government goes with advisor. Separately searching for adviser and advisor on individual websites for well-known institutions in the US and the UK, we can approximate which spelling is more common. As an indicator, let’s use the ratio of advisor-results to adviser-results: the “o/e ratio”. An o/e ratio greater than one means that advisor returns more results; less than one and adviser is more commonly found.
UK university websites use each spelling about equally. US university websites go heavy on advisor. (One exception is my PhD university, RPI, which has a low o/e ratio that approaches the “NY Times abyss.”)
Not only is the debate not clearly split on geographical lines, there may not even be a debate. Maybe it’s just a semantic stubbornness that creates a misplaced superiority and leaves the rest of us confused. This guy falls uncomfortably hard on one side of the advisor/adviser debate.
And this guy falls comfortably on both sides of the debate. A true Quantum Twitterer.
advisor is a title, and adviser is anyone else who advises and is not already an advisor.
The data backs up this claim. By selecting only capitalized Advisor and Adviser, the -or spelling has already passed -er. To capitalize Advisor indicates a title, and its dominance compared to Adviser suggests that there is indeed a distinction, albeit tiny.
So, how does 2014 tech society handle ambiguity?
Wikipedia has a link for adviser but shows the page for advisor. A great side effect of having separate pages with the same content is the possibility to compare page views for each spelling. For all of 2013, the advisor page received 100,361 page views, while adviser 4,165. See the chart below for the 2008-2014 weekly page views.
What’s even more interesting for Wikipedia is that there are many specific types advisers with their own page, and many advisors with their own page. And some types of advisors redirect to advisers. And other types of advisers redirect to advisors. For example, “Financial advisor” redirects to “Financial adviser” and “Technical adviser” redirects to “Technical advisor”. This is despite the fact that “Financial advisor” is a much more prevalent phrase than “Financial adviser”, and “Technical adviser” is a more prevalent phrase than “Technical advisor”, as we’ll soon discover.
Both the standard dictionaries, the Oxford English Dictionary and the American Merriam-Webster, have links for advisor but show the main spelling as adviser. Interestingly, if you search directly at m-w.com, the o/e-neutral page for advise comes up, which seems to equally frustrate everyone (e.g. the comments).
Thesaurus.com show many more results for advisor. The first synonym listed is another victim of careless machine learning and is blazingly incorrect—adviser! Then, to make matters worse, the adviser page doesn’t show advisor as a synonym.
Tripadviser.com has a redirect to tripadvisor.com. And good thing for the redirect — there are still 10,400 Google results for tripadviser.com. (It’s common for large sites to buy domains of common spelling errors and typos, i.e. twittter.com). Here is an example of one Twitter user fighting for his/her semantic preference. Or maybe it was that pesky autocorrect, also guilty of forgetting the apostrophe in “cant”.
But let’s dig deeper and see what the data says.
(please visit the Code Section for further description & links to codes)
On Twitter, it’s not straightforward to count the number of tweets that are found with a search term. Doing so required some mini-mining with the Twitter API and a snippet of Python code. I downloaded tweets that have either advisor or adviser and then counted each separately. Going back in time one week, there were 50,165 times that the case-insensitive advisor was used, compared to 23,016 for adviser, corresponding to an o/e ratio of 2.2. Counting only instances of capitalized Advisor and Adviser, their counts were 25,383 and 7,825, respectively. With an increased o/e ratio of 3.2, the predominance of capitalized Advisor indicates a title or a role, as described above. By putting the tweets with public geo-location on the world map, with advisor in blue and adviser in red, nothing really stands out. But it’s sometimes fun to click and see what people around the world are saying about their advisors and their advisers?
Word clouds, which are normally pretty useless, show some distinctions between the spellings. Employment terms such as job, sales, financial, customer, and service are commonly found in tweets containing advisor. Tweets with adviser include more political terms, such as legal, national, security, and president. I guess word clouds are still pretty useless.
Finally, tweets are filtered by emotion, literally by adding a smiley or frowny face to the Twitter search string. Happy tweets over one week were well-balanced: 477 advisors and 593 advisers (0.8 o/e ratio, which is much fewer non-filtered tweets). In sad tweets, advisor showed up 48 times, compared to 157 for adviser (0.3 o/e ratio). It’s not a huge number of tweets and I don’t want to make any dramatic conclusions, such as claiming that “advisers are more polarizing than advisors.”
Et tu, Goo-gle?
At the time of writing, Google approximated 119,000,000 results for advisor and only 20,500,000 for adviser. That’s a very large o/e ratio of 5.8. Despite that lopsidedness, only adviser brings up a definition box. (For a short discussion about using approximate Google search results, see Appendix 1.)
By using approximate results from Google search, it’s possible to build an understanding of which spelling is preferred in certain contexts. For example, searching “Financial Advisor” with quotes returns 5,120,000 results and “Financial Adviser” returns 1,480,000. (For the methodology of term collection, see Appendix 2.) In addition to well known roles like “Academic Advisor”, I’ve added some general adjectives, such as “trusted” and “good”.
The results are… complicated. “Financial advisor” is the most common. “Afraid adviser” is the least. “Stock adviso/er” has the highest o/e ratio (185.7). Categories consistently with more advisors than advisers are academic, business and health. Despite the US Government’s preference for advisor, most political advisers use the -er spelling.
Search phrases with an o/e ratio of less than 1.0 are most interesting because there are, in general, fewer advisers than advisors online. Even more interesting is if those advisers include American-specific spellings, because adviser is often considered the British spelling. And even more interesting is if that type of adviser includes the American spelling. The double-American spelling of “jewelry advisor” has more results than the British/American mix-match of “jewellery advisor”, which has more results than the double-British “jewellery adviser”. All of those are dwarfed by “Jewelry adviser”, by far the most popular choice, and because of that dominance, “jewelry adviso/er” has the second lowest o/e ratio.
All the British English spellings land above the 1.0 o/e ratio line.
Maybe there is no answer.
Where do we go from here?
I took a look in my past in order to see where to go in my future. In my personal and professional life, after scanning 10 years, 8gb and 56,019 sent emails, the count is:
adviser, 70 times, in 57 emails
advisor, 48 times, in 40 emails
…and I’ve never posted a review on tripadvisor.com OR tripadviser.com. As a reference, I’ve used the word ‘the’ 113,517 times in 15,264 emails.
Relax. Both spellings are fine for all purposes. The general rule says you stick with one spelling in one document. But who owns that rule, anyway? I think it’s time to fully embrace the ambiguity. I want to live in a world where the spell-checker randomly reassigns your already correct spelling to another correct spelling. We’ll call it the Quantum Spellcorrector and we’ll laugh about back-in-the-day when it mattered.
Although extremely approximate and fluctuating, many language learners use the number of Google results to determine what sounds more natural. For example, in German, gern and gerne are interchangeable, and mean “gladly”. Despite that, native speakers will often converge on one spelling in one certain context.
Interestingly, Google’s N-grams tool indicate that gerne has recently passed gern in popularity.
Which is good, because I prefer gerne. I think it makes my German sound Italian. Ich hätte gerne eine pizza.
Now, who knows the literary reason for the bump in the popularity of gerne in the late 1940s? I’ll post any good guesses here:
- Discussion based on my question to german.stackexchange.com. No concrete answer yet, but a link to a similar finding for another set of words. It seems to be a coincidence.
In doing the research for this article, I learned at least two interesting things:
- It’s possible to do a Google search with “Autostart”, instead of autocomplete.
- Ejaculation Advisor is a thing.
You are encouraged to use and improve.
- To get Wikipedia page views, there is a handy webpage that also has an API. I wrote a small python script to collect data for my target pages.
- For anything with the Twitter API, you will need to register as a developer and set up your “app”. You can start here. Once you have all your access codes, take a look at python-twitter library. This code snippet is what I use to download tweets.
- To aggregate the count of Google search results, here is my code snippet. There is a good chance that you will get your IP blocked, so be patient and run smaller lists.
- To download your gmail mailbox, follow these steps. Then, to analyze the .mbox file, check out this code snippet.
Made it this far? Why not say Hi! on Twitter? @philshem