Social penetration theory: the onion layers a chat keeps
Social penetration theory describes closeness as disclosure widening in breadth and deepening in depth. A long chat keeps that progression, and its retreats, in order.
Social penetration theory, set out by Irwin Altman and Dalmas Taylor in their 1973 book Social Penetration, describes closeness as self-disclosure moving through layers, like an onion, along two measures: breadth, the range of topics people talk about, and depth, how far into a topic they go.
The theory is about talk, and most talk disappears as it happens. A long text conversation is an exception. It keeps every topic that came up and every message that went further than the last, in the order they were written, which makes it one of the few places where the widening and deepening the theory describes can be looked at afterwards in the participants’ own words.
Breadth and depth, the two directions into the onion
The book is catalogued as Social penetration: the development of interpersonal relationships, by Irwin Altman and Dalmas A. Taylor, published in New York by Holt, Rinehart and Winston in 1973, according to its Open Library catalogue record. A research overview by Narissra Punyanunt-Carter, published by EBSCO in 2019, describes the central image: Altman and Taylor compared a person to an onion with many layers, an outer layer anyone can see and inner layers that are only revealed as a relationship develops.
The same overview separates the two directions in which the layers open. Breadth is which topics people are willing to talk about with each other, and a relationship becomes more intimate as more and more varied subjects become open to discussion. Depth is how far a person is willing to go into a given topic. Its example is a person who will name a favourite film to someone new but is not yet comfortable describing what the film made them feel, until the other person reciprocates or the two grow closer.
Those are two different movements, and they need not happen together. A conversation can range over a dozen subjects and stay on the surface of all of them, or stay on one subject and go a long way down.
Four stages of disclosure, and a fifth that runs backwards
The EBSCO overview lists the stages the model describes. The orientation stage is superficial talk, the weather and other safe matters, with most people following social norms and steering away from anything that could offend. The exploratory affective stage is where people share feelings about things like favourite foods or films, and the overview notes that many friendships stay there. The affective stage opens more private topics, and with it more willingness to criticise, which can bring disagreement. The stable stage is the deepest, where people share their most intimate details.
There is also a possible fifth stage, depenetration, in which disclosures decrease. The overview describes it as happening when a relationship begins to fail or when the costs of intimate disclosure come to outweigh the benefits, and says it may end a relationship entirely or simply withdraw certain intimate details. Altman and Taylor proposed, in the overview’s words, that relationships succeed when disclosures are rewarding and fail when they prove costly.
Why a chat export holds the layers in order
Social penetration theory was written about relationships in general, not about text messages. What makes a chat relevant is not that the theory was built for it, but that a chat preserves the material the theory talks about. A spoken conversation leaves no transcript unless someone makes one. A written conversation is its own transcript.
Read from the top, a long chat is a sequence of disclosures by two people, each placed after the one before it. Breadth shows up as the set of subjects the conversation touches. Depth shows up inside a subject, in the difference between the first time it was mentioned and the later message that went further. Neither needs interpretation to be seen; they are on the page, in order.
Later research has carried the model into writing online. A study of 1,027 Taiwanese bloggers by Jih-Hsin Tang and Cheng-Chung Wang, indexed on PubMed, described its own design as based on social penetration theory, sorted what the bloggers disclosed into nine topics, and measured the depth and width of disclosure to three audiences: an online audience, a best friend and parents. It found that the bloggers disclosed most to best friends, then parents, then online audiences, in both depth and width. Those are answers to one online survey of bloggers in Taiwan, about blogs rather than private chats, and they say nothing directly about any one conversation.
Topics that widen over months of messages
If a chat between two people who have just met follows the model, its early stretch is narrow in the way the orientation stage describes: a few safe subjects, revisited. In the model’s terms, breadth is what widens across the following weeks or months, as subjects such as work, family, plans, past events and opinions each make a first appearance. Scrolling an export, the first mention of a subject is easy to find, and so is the point where the list of things the two people talk about stopped growing.
Depth is slower to spot because it lives inside topics rather than between them. A subject that first appears as a passing fact can come back months later as a long message about what it meant to the writer. The film example from the overview is the pattern in miniature: the title first, the feeling later, once the other person has offered something of their own.
That last step, one disclosure answered with another, has been tested in text. In an experiment published by L. Crystal Jiang, Natalya Bazarova and Jeffrey Hancock in Communication Research in 2013, participants talked with a confederate who made either intimate or non intimate disclosures, either through text based computer mediated communication or face to face. Participants in the text condition perceived the confederate’s disclosures as more intimate, and reciprocated with more intimate disclosures of their own, than those face to face. That is a laboratory result about conversations with a stranger, not a measurement of how any real chat develops. The broader question of why text can feel closer than talk has its own page, on the hyperpersonal model.
When the messages pull back to the outer layers
The fifth stage is the part of the model a chat preserves most plainly, because retreat leaves a gap where something used to be. Depenetration in a written record looks like topics that used to recur and stop, long messages replaced by short ones, and questions that used to get a paragraph getting a line. The overview’s description allows for two versions: an ending, or a withdrawal of certain intimate details while the relationship continues.
The record shows that the pattern changed. It does not show why. A conversation can narrow because something went wrong, because the talk moved to a phone call or a kitchen table, because one person was busy for a season, or because a subject was settled and no longer needed words. Much of a relationship may never pass through the chat at all, which is why the same caution runs through what a chat analysis cannot tell you about someone.
Where the onion model and the record both fall short
The model has known weak points. The EBSCO overview records that Altman and Taylor’s original model set out a linear progression from stage to stage, and that other researchers have argued the process is dialectical instead, with relationships moving back and forth and depenetration occurring repeatedly between new periods of penetration. It also notes that scholars have pointed to gender, ethnicity and race as influences on the process and on the experiments used to test it, and that much research has looked at how technology and social media disrupt the traditional model.
A chat export has weak points of its own as a record of layers. It holds what was typed in that one channel and nothing said anywhere else. It cannot separate a topic that never came up from a topic that was covered in person. It records what each person chose to write, which is a disclosure about the writer only to the extent that the writer meant it. Breadth and depth are also the theory’s terms, not quantities printed in the file, so any count of topics or any judgement of how deep a message goes is a reading of the record rather than something stored in it.
Choosing whose layers and which months Portraitor reads
Portraitor’s home page describes a four step flow: the conversation, who, when, and payment. At the who step it asks which person in the conversation to analyse, with the instruction to pick the one person you want a portrait of, and says that their messages are read while everyone else stays context. At the when step it offers process all messages, the last 12 months, last 6 months, last 3 months, last month, or a custom range dragged across the conversation. Both choices come before payment.
In social penetration terms those two settings decide whose side of the layering is read and over which stretch of it. A custom range could cover the months when the topics were widening, or the later stretch where they narrowed. The published pages do not say that Portraitor measures breadth, depth or the stages of the onion model, and this page does not claim it does. The detail of the time window is on analysing only part of a chat, and the one person choice, including picking yourself, is on analysing your own messages. The tool itself is at portraitor.ai.
Petronio’s communication privacy management theory approaches the same material from the other side, reading a disclosure as a boundary being moved rather than a layer being peeled back: communication privacy management theory sets out the rules it says people negotiate.
Frequently asked questions
Frequently asked questions
What is social penetration theory in simple terms?
It is the idea, set out by Irwin Altman and Dalmas Taylor in 1973, that people become close by disclosing more about themselves over time, moving from outer layers of a personality towards inner ones, like peeling an onion. The EBSCO overview describes it this way.
What is the difference between breadth and depth in social penetration theory?
Breadth is the range of topics two people are willing to talk about, and depth is how far into any one topic they go. A conversation can be broad and shallow, or narrow and deep, and the theory counts both as part of becoming closer.
What are the stages of social penetration theory?
The stages described in the EBSCO overview are orientation, exploratory affective, affective and stable, with a possible fifth stage, depenetration, in which disclosure decreases. Critics cited in the same overview argue that real relationships move back and forth between stages rather than through them in order.
Does social penetration theory apply to texting?
The theory was written about relationships in general, and later researchers have applied it to writing online, including a PubMed indexed survey of Taiwanese bloggers that measured the depth and width of their disclosure. None of that tells you how a particular text conversation developed, but a chat does keep the written part of it in order.
What is depenetration?
Depenetration is the stage in which people disclose less, either as a relationship ends or as certain intimate details are withdrawn while it continues. In a chat it can appear as topics that stop recurring and messages that get shorter, though the record cannot say why that happened.
Disclosure moving inward and then back out is one way to read a long record. Relational dialectics theory: a tension that never settles reads the same material a different way, as competing discourses that never resolve.
The onion model is a theory of talk, and a long chat is one of the few places that talk is still on the page.
Social penetration theory describes what a person discloses, layer by layer. How the reply to that disclosure was received is a separate quantity, rated by the discloser rather than read off the thread, and perceived partner responsiveness, and what a chat log misses sets out what a chat export keeps of it.
Your conversations, read the way you would read someone else’s.
Portraitor takes a chat export you already have, anonymises it in your browser, and returns a structured reading of how you come across. Identifying details are removed in your browser before anything is sent.
See how it works →- Altman, I. and Taylor, D. A. (1973). Social penetration: the development of interpersonal relationships. Holt, Rinehart and Winston. Open Library catalogue record.
- Punyanunt-Carter, N. (2019). Social penetration theory. EBSCO Research Starters.
- Tang, J.-H. and Wang, C.-C. (2012). Self-disclosure among bloggers: re-examination of social penetration theory. Cyberpsychology, Behavior, and Social Networking. PubMed 22489546.
- Jiang, L. C., Bazarova, N. N. and Hancock, J. T. (2013). From perception to behavior: disclosure reciprocity and the intensification of intimacy in computer-mediated communication. Communication Research, 40(1).
- Portraitor home page, the four step flow.