Some of us might remember a time before the answer to everything was a (flippant) “Google it”. You had to consider where you might find the information you wanted, physically retrieve that source, hope it had a good index, and page through it. And if you were wrong, the process started all over again. Having powerful computer search tools certainly makes researching easier and faster, but it doesn’t always make it better, more efficient, or more targeted. Search engines like Google in particular can be scattershot, returning popular results rather than academic ones. In this guest post, we learn how to target your searches to get the best of both worlds: the precision of a reference library plus the speed of the internet.
The Basics of Boolean Search
Being able to use booleans to tell a search what you do and do not want is a highly effective way to find the sources you want for your own research. A boolean at its core is a logical proposition; that is, it can be only either true or false. In the context of searching, booleans are an easy way for computers to decide if an item in a database contains information you want. Does this journal article mention Herodotus? True. Was it published after 1990? False. Below is a basic overview of boolean searching, from the very basics to complex nested booleans. Other aspects of searching will require further posts.
Simple single term/phrase searching is quite straightforward and something with which most people are acquainted. For the purpose of providing a concrete visual example, I will be using the EBSCOHost interface for L’année Philologique as it is provided by the University of Georgia Library System.
Putting in a search for “Homer” will result in all sources in this database for which the word “Homer” appears in any field (Author, Subject, Text, etc.) in the item.
Most boolean searches allow searchers to narrow down their queries so they’re only looking at certain aspects of the items in the database. While some require you to choose a field, many will default to searching through every field if you do not specify one.
The real power of boolean searching comes in when one needs to do a search that involves multiple topics. Different sites have different ways of formatting their more complex boolean searches, but generally there are three main concepts: AND, OR, and NOT.
If I want to search for all items in L’année Philologique in which “Homer” is listed as an Ancient Author and the word “dog” appears anywhere in the item, I can use the AND operator to locate only items that contain both Homer as an Author and dog as a word somewhere in the item.
When a database does a boolean search, it takes each item within it and asks, “Does this item meet the criterion I have been given?”. If the answer is Yes/True, then it will give the result to you since it might be what you wanted to find; it then goes on to the next item. If the answer is No/False, it rejects it and moves on. When the database is given two criteria connected by an AND, it only returns items for which the answers to both criteria are True. If either one (or both) are False – that is, one criterion is not met – then the item is rejected. In computer science, often the AND operator is referred to by the “&” symbol.
The OR operation is good for when there are multiple terms used to refer to the topic or other situations in which you want to look for items that fit either criterion A or criterion B (or even C or D or E…). Most OR searches are not exclusive, meaning that OR really means “and/or”. That is, if a result happens to fit both criteria, it will still be accepted and presented to you. In computer science, the “|” symbol represents the OR operator. In the example above, the search is asking for all items in the database that are about Roman epic OR Greek epic (or both). The OR operation is quite useful if you are intending to compare/contrast two different topics (e.g. Roman Epic vs. Greek Epic).
The NOT operation is possibly the least utilized boolean operator, but it can be very useful when you want to avoid a specific subtopic in your research. For example, if you’re writing an article on Vergil but you want to find sources on him that don’t reference the Aeneid, you can use the NOT operator to tell the database to reject anything with the word “Aeneid” in it. It can also be useful is there is an article or book that keep coming up in your searches that does not pertain to your specific area of research. Using NOT to “blacklist” this item’s author, title, etc. will prevent it from showing up in your future searches. NOT is more or less the logical opposite of the standard search, because instead of the database asking “does this item have X,” it’s asking “does this item NOT have X”. And if the answer to that second question is Yes/True, then it will return that result (assuming any other criteria are also met). In computer science, the “^” symbol is usually used to refer to the NOT operation.
If you are still a bit confused about what AND, OR, and NOT mean or when to use them, this graphic from Ohio University Libraries uses Venn diagrams to illustrate the three boolean
Know When It’s Nested and When It’s Not
It;s important to note that unless multiple criteria are intentionally nested (which I will discuss shortly), they are not nested. Take the example above: you are searching for midwifery or childbirth and ancient Greece, not midwifery or (childbirth and ancient Greece). Your search as it stands will return results with “midwifery” somewhere in it and may or may not also use the term “childbirth” somewhere, and will also definitely also have “ancient Greece” in it. [Ed’s note: if you remember how the order of operations in math works, the parentheses function the same way here. 3+4×5 =/= (3+4)x5!]
Using the three boolean concepts of AND, OR, and NOT, one can combine multiple searches into one more specific (and more complex) search.
In the search above I have nested a search for either the term “dog” or “hund” (the German word for dog) within a larger search for both Homer AND my nested boolean. Since the computer is simply searching for the string of letters “d”, “o”, and “g” in that order, it is not going to recognize the word “hund” as something in which I may also be interested. To help the computer know I am also interested in German-language articles on this subject, I need to include a statement that searches for either “dog” or “hund”.
Nested booleans (which need not always be OR operations, by the way) can also be useful when your topic of interest may be referred to in a variety of ways. Say you are researching pregnancy in the ancient world. Works on childbirth in the ancient world would still probably be useful to you, so you will want to search for works on either “childbirth” or “pregnancy”. Works on your topic may also not specifically say “ancient world,” but may instead refer to a more specific location. By including a search for “ancient Greece” OR “ancient Rome,” you can get the most results that pertain to my area of research.
Nested queries can be represented as parentheticals nested within parentheticals.
Ex. 1: Homer AND (dog OR hund)
Ex. 2: (childbirth | pregnancy) & (ancient greece | ancient rome)
Below is a fun example of a very complex nested boolean search. Try to figure out what is being searched!
((sculpture | carving) & relief) & ((baroque | mannerist) ^ renaissance)
Victoria Hiten is studying Computer Science and Classical Languages at the University of Georgia. This post stems from a project she undertook at the CHS Information Fluency Workshop.